Safety vulnerability ID: 59334
The information on this page was manually curated by our Cybersecurity Intelligence Team.
H2o 3.44.0.1 updates its MAVEN dependency 'org.codehaus.jettison:jettison' to '1.5.4' to fix CVE-2022-40150.
The dependency is included in /h2o/backend/bin/h2o.jar
Latest version: 3.46.0.6
H2O, Fast Scalable Machine Learning, for python
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-3.42.0/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-3.42.0/1/index.html</a>
Bug
- [[15423]](https://github.com/h2oai/h2o-3/issues/15423) - Fixed Infogram cross-validation with weights.
- [[15482]](https://github.com/h2oai/h2o-3/issues/15482) - Updated R package maintainer.
- [[15461]](https://github.com/h2oai/h2o-3/issues/15461) - Fixed leaks in GLM’s Negative Binomial estimation.
Improvement
- [[6843]](https://github.com/h2oai/h2o-3/issues/6843) - Changed `warning` tag to `info` tag when weights are not provided during validation/test dataset scoring when weights are present in training.
- [[6828]](https://github.com/h2oai/h2o-3/issues/6828) - Removed support for Python 2.7 and 3.5.
- [[6813]](https://github.com/h2oai/h2o-3/issues/6813) - Upgraded the default parquet library to 1.12.3 for standalone jar.
- [[7630]](https://github.com/h2oai/h2o-3/issues/7630) - Upgraded XGBoost to version 1.6.1.
New Feature
- [[6548]](https://github.com/h2oai/h2o-3/issues/6548) - Implemented AIC metric for all GLM model families.
- [[6880]](https://github.com/h2oai/h2o-3/issues/6880) - Implemented Tweedie variance power maximum likelihood estimation for GLM.
- [[6943]](https://github.com/h2oai/h2o-3/issues/6943) - Added ability to convert H2OAssembly to a MOJO2 artifact.
- [[7008]](https://github.com/h2oai/h2o-3/issues/7008) - Implemented new Decision Tree algorithm.
Docs
- [[15474]](https://github.com/h2oai/h2o-3/issues/15474) - Added link to AutoML Wave app from AutoML user guide.
- [[15550]](https://github.com/h2oai/h2o-3/issues/15550) - Added documentation on H2OAssembly to MOJO 2 export functionality.
- [[15602]](https://github.com/h2oai/h2o-3/issues/15602) - Added algorithm page in user guide for new Decision Tree algorithm.
- [[15529]](https://github.com/h2oai/h2o-3/issues/15529) - Added AIC metric support for all GLM families to GLM user guide page and GLM booklet.
- [[15466]](https://github.com/h2oai/h2o-3/issues/15466) - Updated authors and editors for GLM booklet.
- [[6884]](https://github.com/h2oai/h2o-3/issues/6884) - Added documentation on Tweedie variance power maximum likelihood estimation to GLM booklet and user guide.
- [[7200]](https://github.com/h2oai/h2o-3/issues/7200) - Improved user guide documentation for Generalized Additive Models algorithm.
Security
- [[15594]](https://github.com/h2oai/h2o-3/issues/15594) - Addressed CVE-2023-2976 in h2o-steam.jar.
- [[15548]](https://github.com/h2oai/h2o-3/issues/15548) - Addressed CVE-2020-29582 in h2o-steam.jar.
- [[15546]](https://github.com/h2oai/h2o-3/issues/15546) - Addressed CVE-2023-26048 and CVE-2023-26049 by upgrading Jetty for minimal and steam jar.
- [[15540]](https://github.com/h2oai/h2o-3/issues/15540) - Addressed PRISMA-2023-0067 in h2o-steam.jar.
- [[6827]](https://github.com/h2oai/h2o-3/issues/6827) - Addressed CVE-2023-1436, CVE-2022-45693, CVE-2022-45685, and CVE-2022-40150 by upgrading org.codehaus.jettison:jettison in h2o-steam.jar.
Kurka (3.40.0.4) - 4/28/2023
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/4/index.html</a>
Bug
- [[6758]](https://github.com/h2oai/h2o-3/issues/6758) - Fixed the deprecation warning thrown for Python 2.7 and 3.5.
Improvement
- [[6756]](https://github.com/h2oai/h2o-3/issues/6756) - Added official support for Python 3.9.
Docs
- [[6759]](https://github.com/h2oai/h2o-3/issues/6759) - Removed mention of support for Python 2.7 and Python 3.5 from documentation.
- [[7600]](https://github.com/h2oai/h2o-3/issues/7600) - Reorganized supervised and unsupervised algorithm parameters by algorithm-specific, common, and shared-tree (for tree-based algorithms). Updated parameter descriptions for all supervised and unsupervised algorithms. Shifted all shared GLM family parameters to the GLM algorithm page.
Security
- [[6732]](https://github.com/h2oai/h2o-3/issues/6732) - Addressed CVE-2023-1370 by removing the vulnerability from h2o-steam.jar.
Kurka (3.40.0.3) - 4/4/2023
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/3/index.html</a>
Improvement
- [[6763]](https://github.com/h2oai/h2o-3/issues/6763) - Added GAM Knot Locations to Model Output.
- [[6764]](https://github.com/h2oai/h2o-3/issues/6764) - Addressed CVE-2014-125087 in h2o-steam.jar
Engineering Story
- [[6767]](https://github.com/h2oai/h2o-3/issues/6767) - Disabled execution of tests in client mode.
- [[6772]](https://github.com/h2oai/h2o-3/issues/6772) - Deprecated support for Python 2.7 and 3.5.
Docs
- [[6773]](https://github.com/h2oai/h2o-3/issues/6773) - Introduced a page describing MOJO capabilities.
- [[6790]](https://github.com/h2oai/h2o-3/issues/6790) - Updated the DRF documentation page to reflect what dataset is used to calculate the model metric.
- [[6793]](https://github.com/h2oai/h2o-3/issues/6793) - Updated and rearranged the hyper-parameter list in the Grid Search documentation page.
Kurka (3.40.0.2) - 3/9/2023
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/2/index.html</a>
Bug
- [[6818]](https://github.com/h2oai/h2o-3/issues/6818) - Fixed dependency on numpy in Fairness-related code.
- [[6819]](https://github.com/h2oai/h2o-3/issues/6819) - Added ability to debug GBM reproducibility by looking at tree structure with `equal_gbm_model_tree_structure`.
Improvement
- [[6995]](https://github.com/h2oai/h2o-3/issues/6995) - Fixed the deviance computation for GBM Poisson distribution.
New Feature
- [[6777]](https://github.com/h2oai/h2o-3/issues/6777) - Added `save_plot_path` parameter for Fairness plotting allowing you to save plots.
Task
- [[6538]](https://github.com/h2oai/h2o-3/issues/6538) - Implemented incremental MaxRSweep without using sweep vectors.
- [[6799]](https://github.com/h2oai/h2o-3/issues/6799) - Removed duplicate predictors for ModelSelection’s MaxRSweep.
Engineering Story
- [[6776]](https://github.com/h2oai/h2o-3/issues/6776) - Pointed MLOps integration to internal.dedicated environment.
Docs
- [[6501]](https://github.com/h2oai/h2o-3/issues/6501) - Added warning that `max_runtime_secs` cannot always produce reproducible models.
- [[6503]](https://github.com/h2oai/h2o-3/issues/6503) - Added example for how to save a file as a parquet.
- [[6811]](https://github.com/h2oai/h2o-3/issues/6811) - Added example for how to connect to an H2O cluster by name.
- [[6887]](https://github.com/h2oai/h2o-3/issues/6887) - Added information on the implementation of the `eval_metric` for XGBoost.
Kurka (3.40.0.1) - 2/8/2023
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zz_kurka/1/index.html</a>
Bug
- [[6845]](https://github.com/h2oai/h2o-3/issues/6845) - Improved GLM negative binomial calculation time.
- [[6882]](https://github.com/h2oai/h2o-3/issues/6882) - Cleaned up COLLATE field in the description of the R package by allowing Roxygen2 to generate the COLLATE field.
- [[6891]](https://github.com/h2oai/h2o-3/issues/6891) - Changed the exceptions in Stacked Ensembles checks to ModelBuilder warnings.
- [[7090]](https://github.com/h2oai/h2o-3/issues/7090) - Fixed GLM ignoring time budget when trained using cross-validation in AutoML.
- [[7132]](https://github.com/h2oai/h2o-3/issues/7132) - Fixed incorrect actual `ntrees` value reported in tree-based models.
Improvement
- [[6805]](https://github.com/h2oai/h2o-3/issues/6805) - Increased speed of XGBoost scoring on wide datasets.
- [[6864]](https://github.com/h2oai/h2o-3/issues/6864) - Updated error message for when a user specifies the wrong cluster when connecting to a running H2O instance.
- [[6886]](https://github.com/h2oai/h2o-3/issues/6886) - Improved memory usage in creation of parse-response for wide datasets.
- [[6893]](https://github.com/h2oai/h2o-3/issues/6893) - Increased testing speed by adding ability to train XGBoost cross-validation models concurrently on the same GPU.
- [[6900]](https://github.com/h2oai/h2o-3/issues/6900) - Added ability to score `eval_metric` on validation datasets for XGBoost.
- [[6901]](https://github.com/h2oai/h2o-3/issues/6901) - Added notebook demonstrating `eval_metric` for XGBoost.
- [[6902]](https://github.com/h2oai/h2o-3/issues/6902) - Increased XGBoost model training speed by disabling H2O scoring to rely solely on `eval_metric`.
- [[6910]](https://github.com/h2oai/h2o-3/issues/6910) - Updated to Java 17 from Java 11/openjdk in H2O docker images.
- [[7294]](https://github.com/h2oai/h2o-3/issues/7294) - Updated warning message for when H2O version is outdated.
- [[7598]](https://github.com/h2oai/h2o-3/issues/7598) - Introduced a better format for storing default, input, and actual parameters in H2O model objects for R by using `params` slots.
- [[7835]](https://github.com/h2oai/h2o-3/issues/7835) - Added `model_summary` to Stacked Ensembles.
- [[7980]](https://github.com/h2oai/h2o-3/issues/7980) - Moved StackedEnsembleModel::checkAndInheritModelProperties to StackedEnsemble class.
New Feature
- [[6858]](https://github.com/h2oai/h2o-3/issues/6858) - Added ability to publish models to MLOps via Python API.
- [[7009]](https://github.com/h2oai/h2o-3/issues/7009) - Added ability to grid over Infogram.
- [[7044]](https://github.com/h2oai/h2o-3/issues/7044) - Implemented Regression Influence Diagnostics for GLM.
- [[7045]](https://github.com/h2oai/h2o-3/issues/7045) - Enhanced GBM procedures to output which records are used for each tree.
- [[7537]](https://github.com/h2oai/h2o-3/issues/7537) - Added learning curve plot to H2O’s Explainability.
Task
- [[6802]](https://github.com/h2oai/h2o-3/issues/6802) - Added `negative_log_likelihood` and `average_objective` accessor functions in R and Python for GLM.
- [[7088]](https://github.com/h2oai/h2o-3/issues/7088) - Limited the number of iterations when training the final GLM model after cross-validation.
Technical Task
- [[6898]](https://github.com/h2oai/h2o-3/issues/6898) - Added support for scoring `eval_metric` on a validation set for external XGBoost cluster.
- [[6899]](https://github.com/h2oai/h2o-3/issues/6899) - Added support for scoring `eval_metric` on a validation set for internal XGBoost cluster.
- [[7012]](https://github.com/h2oai/h2o-3/issues/7012) - Implemented GLM dispersion estimation parameter using maximum likelihood method for the negative binomial family.
Docs
- [[6820]](https://github.com/h2oai/h2o-3/issues/6820) - Highlighted information about how rebalancing makes reproducibility impossible.
- [[6815]](https://github.com/h2oai/h2o-3/issues/6815) - Added documentation on the `negative_log_likelihood` and `average_objective` accessor functions.
- [[6816]](https://github.com/h2oai/h2o-3/issues/6816) - Added information on GLM dispersion estimation using maximum likelihood method for the negative binomial family.
- [[6821]](https://github.com/h2oai/h2o-3/issues/6821) - Added documentation on Regression Influence Diagnostics for GLM.
- [[6803]](https://github.com/h2oai/h2o-3/issues/6803) - Fixed non-functional data paths in code examples throughout the user guide.
- [[6804]](https://github.com/h2oai/h2o-3/issues/6804) - Added information on the `row_to_tree_assignment` function.
- [[6807]](https://github.com/h2oai/h2o-3/issues/6807) - Added documentation on using H2O with Apple M1 chip.
- [[6808]](https://github.com/h2oai/h2o-3/issues/6808) - Added information on `init` parameter being skipped due to `estimate_k=True` for K-Means.
Zygmund (3.38.0.4) - 1/5/2023
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/4/index.html</a>
Bug
- [[6851]](https://github.com/h2oai/h2o-3/issues/6851) - Fixed error in SHAP values report for DRF.
- [[6865]](https://github.com/h2oai/h2o-3/issues/6865) - Fixed a ModelSelection replacement error stopping too early and implemented incremental forward step and incremental replacement step for numerical predictors.
Task
- [[6852]](https://github.com/h2oai/h2o-3/issues/6852) - Resolved hyperparameters amongst the algorithms.
- [[6857]](https://github.com/h2oai/h2o-3/issues/6857) - Removed redundant predictors found in `mode=“backward”` for ModelSelection.
Engineering Story
- [[6846]](https://github.com/h2oai/h2o-3/issues/6846) - Renamed the docker image `h2o-steam-k8s` to `h2o-open-source-k8s-minimal`.
Docs
- [[6800]](https://github.com/h2oai/h2o-3/issues/6800) - Updated download page by adding options for steam jar and python client without h2o backend.
- [[6849]](https://github.com/h2oai/h2o-3/issues/6849) - Fixed log likelihood of negative binomial for GLM.
- [[6855]](https://github.com/h2oai/h2o-3/issues/6855) - Added how users can force an unsupported Java version.
- [[6856]](https://github.com/h2oai/h2o-3/issues/6856) - Fixed broken links on the H2O Release page.
- [[6860]](https://github.com/h2oai/h2o-3/issues/6860) - Added information on how Isolation Forest and Extended Isolation Forest handle missing values.
- [[6862]](https://github.com/h2oai/h2o-3/issues/6862) - Fixed typos and made examples work on performance-and-prediction.html.
- [[6863]](https://github.com/h2oai/h2o-3/issues/6863) - Removed outdated roadmap from Readme file.
Security
- [[6794]](https://github.com/h2oai/h2o-3/issues/6794) - Addressed CVE-2022-3509 by upgrading `google-cloud-storage`.
Zygmund (3.38.0.3) - 11/23/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/3/index.html</a>
Bug
- [[6877]](https://github.com/h2oai/h2o-3/issues/6877) - Enforced DkvClassLoader while accessing Python resources through JythonCFuncLoader.
- [[6878]](https://github.com/h2oai/h2o-3/issues/6878) - Closed open file descriptors from H2OConnection.
- [[6871]](https://github.com/h2oai/h2o-3/issues/6871) - Fixed incorrect value indicator for a partial dependence plot for its current row.
- [[6873]](https://github.com/h2oai/h2o-3/issues/6873) - Fixed GBM model with `interaction_constraints` only building single-depth trees.
- [[6897]](https://github.com/h2oai/h2o-3/issues/6897) - Fixed slow estimator validation when training model with wide datasets.
- [[6907]](https://github.com/h2oai/h2o-3/issues/6907) - Fixed GAM failure when `numknots=2` for I-spline.
Task
- [[6896]](https://github.com/h2oai/h2o-3/issues/6896) - Ensured non-negative will not overwrite `splines_non_negative` for GAM I-spline.
- [[6921]](https://github.com/h2oai/h2o-3/issues/6921) - Implemented p-value calculation for GLM with regularization.
- [[6925]](https://github.com/h2oai/h2o-3/issues/6925) - Verified the minimum number of knots each spline type can support for GAM.
- [[6926]](https://github.com/h2oai/h2o-3/issues/6926) - Implemented normal (non-monotonic) splines that can support any degrees.
Docs
- [[6874]](https://github.com/h2oai/h2o-3/issues/6874) - Updated `compute_p_value` documentation for GLM and GAM to reflect that p-values and z-values can now be computed with regularization.
- [[6875]](https://github.com/h2oai/h2o-3/issues/6875) - Documented GAM M-splines.
- [[6876]](https://github.com/h2oai/h2o-3/issues/6876) - Updated site logo, favicon, and color scheme to reflect H2O’s brand kit.
- [[6870]](https://github.com/h2oai/h2o-3/issues/6870) - Updated booklet links for GBM, GLM, and Deep Learning on their respective algorithm pages.
- [[6881]](https://github.com/h2oai/h2o-3/issues/6881) - Fixed typo in Model Selection for `build_glm_model` parameter.
- [[6885]](https://github.com/h2oai/h2o-3/issues/6885) - Updated links in the provided bibliography in the FAQ.
- [[6894]](https://github.com/h2oai/h2o-3/issues/6894) - Removed Sparkling Water booklet link from the download page.
- [[6904]](https://github.com/h2oai/h2o-3/issues/6904) - Added optional Python plotting requirement `matplotlib` to the download page.
Zygmund (3.38.0.2) - 10/27/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/2/index.html</a>
Bug
- [[6895]](https://github.com/h2oai/h2o-3/issues/6895) - Fixed H2ODeepLearningEstimator `autoencoder` not working without `y` value.
- [[6911]](https://github.com/h2oai/h2o-3/issues/6911) - Added `libgomp` into docker images thus enabling XGBoost multithreading.
- [[6919]](https://github.com/h2oai/h2o-3/issues/6919) - Stopped throwing warning about jobs not having proper model types when models weren’t even trained.
- [[6928]](https://github.com/h2oai/h2o-3/issues/6928) - Fixed cross validation failure for concurrent sorting.
- [[6930]](https://github.com/h2oai/h2o-3/issues/6930) - Enabled parallelism in cross validation for Isotonic Regression.
Task
- [[6917]](https://github.com/h2oai/h2o-3/issues/6917) - Enabled GAM I-spline to support increasing and decreasing functions.
- [[6925]](https://github.com/h2oai/h2o-3/issues/6925) - Updated the number of knots required for GAM I-splines to be >=2.
- [[6949]](https://github.com/h2oai/h2o-3/issues/6949) - Improved ModelSelection’s `mode=“maxrsweep”` runtime.
Docs
- [[6890]](https://github.com/h2oai/h2o-3/issues/6890) - Added information on ModelSelection’s new `build_glm_model` parameter for `mode=“maxrsweep”`.
- [[6903]](https://github.com/h2oai/h2o-3/issues/6903) - Fixed incorrect header case on ModelSelection and Cox Proportional Hazards algorithm pages in the user guide.
- [[6913]](https://github.com/h2oai/h2o-3/issues/6913) - Added an example to Variable Inflation Factors in the user guide.
- [[6917]](https://github.com/h2oai/h2o-3/issues/6917) - Fixed broken links on the “Welcome to H2O-3” page of the user guide.
- [[6948]](https://github.com/h2oai/h2o-3/issues/6948) - Added model explainability for plotting SHAP to the “Performance and Prediction” page of the user guide.
- [[7442]](https://github.com/h2oai/h2o-3/issues/7442) - Added examples for `varsplits()` and `feature_frequencies()` to Python documentation.
Security
- [[6889]](https://github.com/h2oai/h2o-3/issues/6889) - Addressed CVE-2022-42003 and CVE-2022-42889 security issues through Library upgrades.
Zygmund (3.38.0.1) - 9/19/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zygmund/1/index.html</a>
Bug
- [[6937]](https://github.com/h2oai/h2o-3/issues/6937) - Fixed the sorting of `h2o.make_leaderboard`.
- [[6940]](https://github.com/h2oai/h2o-3/issues/6940) - Fixed H2O dependencies overriding Jetty implementation.
- [[6951]](https://github.com/h2oai/h2o-3/issues/6951) - Fixed Flow’s export Frame throwing an NPE because it doesn’t provide a file type.
- [[6959]](https://github.com/h2oai/h2o-3/issues/6959) - Fixed GLM ordinal generic metrics to provide missing information in the payload.
- [[6960]](https://github.com/h2oai/h2o-3/issues/6960) - Fixed “maxrsweep” NPE in ModelSelection thrown when the replacement step stopped too early.
- [[6961]](https://github.com/h2oai/h2o-3/issues/6961) - Fixed “maxrsweep” replacement bug in ModelSelection by updating the implementation method.
- [[6973]](https://github.com/h2oai/h2o-3/issues/6973) - Fixed unnecessary transformations in the scikit-learn wrapper by using model performance API.
- [[6979]](https://github.com/h2oai/h2o-3/issues/6979) - Fixed upload of big files in Sparkling Water deployment.
- [[6983]](https://github.com/h2oai/h2o-3/issues/6983) - Changed the error message that GLM does not support contributions.
- [[6985]](https://github.com/h2oai/h2o-3/issues/6985) - Fixed QuantilesGlobal histogram type failing in GBM when all columns were categorial.
- [[7002]](https://github.com/h2oai/h2o-3/issues/7002) - Added support for MapR 6.2 to fix the error caused by updating the cluster.
- [[7006]](https://github.com/h2oai/h2o-3/issues/7006) - Fixed large file upload in Python.
- [[7023]](https://github.com/h2oai/h2o-3/issues/7023) - Fixed inability to stop print out of model information in Python.
- [[7056]](https://github.com/h2oai/h2o-3/issues/7056) - Removed `-seed` variable hiding in GAM.
- [[7104]](https://github.com/h2oai/h2o-3/issues/7104) - Updated `h2o.upload_mojo` to also work for POJO.
- [[7432]](https://github.com/h2oai/h2o-3/issues/7432) - Added unsupported operation exception when trying to use SHAP summary plot when building DRF model with `binomial_double_trees`.
- [[8542]](https://github.com/h2oai/h2o-3/issues/8542) - Refactored the rendering logic in the Python client.
- [[10436]](https://github.com/h2oai/h2o-3/issues/10436) - Added xval argument to `h2o.confusionMatrix` in R.
Improvement
- [[6933]](https://github.com/h2oai/h2o-3/issues/6933) - Added support for calibrating an already trained model manually.
- [[6941]](https://github.com/h2oai/h2o-3/issues/6941) - Added support for using Isotonic Regression for model calibration.
- [[6942]](https://github.com/h2oai/h2o-3/issues/6942) - Added ability to S3A allowing it to share the built-in AWS credential providers.
- [[6947]](https://github.com/h2oai/h2o-3/issues/6947) - Improved `configure_s3_using_s3a` allowing it to be usable in any deployment.
- [[6982]](https://github.com/h2oai/h2o-3/issues/6982) - Updated `train_segments` function in R to be independent of camel casing in the algorithm name.
- [[6986]](https://github.com/h2oai/h2o-3/issues/6986) - Improved runtime for QuantilesGlobal histogram by using exact split-points for low-cardinality columns.
- [[6992]](https://github.com/h2oai/h2o-3/issues/6992) - Exposed the Sequential Walker for R/Python and added option to disable early stopping.
- [[7007]](https://github.com/h2oai/h2o-3/issues/7007) - Cleaned up Key API by removing replicas.
- [[7340]](https://github.com/h2oai/h2o-3/issues/7340) - Cleaned up the default output after training a model.
- [[7510]](https://github.com/h2oai/h2o-3/issues/7510) - Exposed calibrated probabilities in `mojo_predict_pandas`.
New Feature
- [[6950]](https://github.com/h2oai/h2o-3/issues/6950) - Simplified the configuration of S3 for Frame exportation.
- [[6984]](https://github.com/h2oai/h2o-3/issues/6984) - Added `train_segments` test for Isolation Forest.
- [[6991]](https://github.com/h2oai/h2o-3/issues/6991) - Added ability to `h2o.no_progress` in R allowing it to accept expressions.
- [[7011]](https://github.com/h2oai/h2o-3/issues/7011) - Implemented dispersion parameter estimation for GLM.
- [[7016]](https://github.com/h2oai/h2o-3/issues/7016) - Added ability to export H2O Frame to a Parquet.
- [[7076]](https://github.com/h2oai/h2o-3/issues/7076) - Added Pareto front plots to AutoML Explain.
- [[7091]](https://github.com/h2oai/h2o-3/issues/7091) - Added “deviance” method to dispersion for calculating p-values.
- [[7093]](https://github.com/h2oai/h2o-3/issues/7093) - Implemented variable inflation factors for GLM.
- [[7192]](https://github.com/h2oai/h2o-3/issues/7192) - Implemented in-training checkpoints for GBM.
- [[8005]](https://github.com/h2oai/h2o-3/issues/8005) - Implemented support for interactions to MOJO for CoxPH.
- [[12152]](https://github.com/h2oai/h2o-3/issues/12152) - Added `h2o.make_leaderboard` function which scores and compares a set of models to AutoML.
Task
- [[6927]](https://github.com/h2oai/h2o-3/issues/6927) - Secured XGBoost connections in multinode environments.
- [[6948]](https://github.com/h2oai/h2o-3/issues/6948) - Added missing added predictor and deleted predictor to the result frame and model summary of ModelSelection.
- [[6952]](https://github.com/h2oai/h2o-3/issues/6952) - Added support allowing you to force GLM to build a null model where the model only returns the coefficients for the intercept.
- [[6953]](https://github.com/h2oai/h2o-3/issues/6953) - Added support allowing GLM `gamma` to fix the dispersion parameter to calculate p-values.
- [[6998]](https://github.com/h2oai/h2o-3/issues/6998) - Implemented “maxr” speedup for Modelselection by introducting “maxrsweep”.
- [[7013]](https://github.com/h2oai/h2o-3/issues/7013) - Implemented dispersion factor estimation using maximum likelihood for GLM gamma family.
Docs
- [[6920]](https://github.com/h2oai/h2o-3/issues/6920) - Added documentation on Isotonic Regression.
- [[6931]](https://github.com/h2oai/h2o-3/issues/6931) - Added variable inflation factors to GLM section of the user guide.
- [[6932]](https://github.com/h2oai/h2o-3/issues/6932) - Added Tweedie dispersion parameter estimation to the GLM section of the user guide.
- [[6934]](https://github.com/h2oai/h2o-3/issues/6934) - Added confusion matrix calculation explanation to performance and prediction.
- [[6939]](https://github.com/h2oai/h2o-3/issues/6939) - Added `get_predictors_removed_per_step()` and `get_predictors_added_per_step()` examples to ModelSelection.
- [[6945]](https://github.com/h2oai/h2o-3/issues/6945) - Added use case section to the welcome page of the user guide.
- [[6987]](https://github.com/h2oai/h2o-3/issues/6987) - Added MOJO import/export information to each algorithm page.
- [[7048]](https://github.com/h2oai/h2o-3/issues/7048) - Added major release blogs to user guide and moved change log to top of the sidebar.
Zumbo (3.36.1.5) - 9/15/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/5/index.html</a>
Security
- Addressed security vulnerability CVE-2021-22569 in the `h2o.jar`.
Zumbo (3.36.1.4) - 8/3/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/4/index.html</a>
Bug
- [[6954]](https://github.com/h2oai/h2o-3/issues/6954) - Disabled `partial_plot` for Uplift DRF temporarily.
- [[6958]](https://github.com/h2oai/h2o-3/issues/6958) - Added support for predicting with Autoencoder when using eigen encoding.
- [[6962]](https://github.com/h2oai/h2o-3/issues/6962) - Fixed XGBoost failure with enabled cross validation in external cluster mode by explicitly starting external XGBoost before cross validation.
Security
- [[6946]](https://github.com/h2oai/h2o-3/issues/6946) - Addressed security vulnerabilities CVE-2021-22573 and CVE-2019-10172 in Steam assembly.
Zumbo (3.36.1.3) - 7/8/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/3/index.html</a>
Bug
- [[6975]](https://github.com/h2oai/h2o-3/issues/6975) - Fixed CoxPH MOJO ignoring offset column.
- [[6976]](https://github.com/h2oai/h2o-3/issues/6976) - Fixed the incorrect predictions from the CoxPH MOJO on categorical columns.
- [[6977]](https://github.com/h2oai/h2o-3/issues/6977) - Fixed the **View** button not working after completing an AutoML job.
- [[6989]](https://github.com/h2oai/h2o-3/issues/6989) - Fixed `num_of_features` not being used in call for `varimp_heatmap()`.
- [[7015]](https://github.com/h2oai/h2o-3/issues/7015) - Fixed GAM’s `fold_column` being treated as a normal column to score for.
- [[7053]](https://github.com/h2oai/h2o-3/issues/7053) - Updated GBM cross validation model summary tables to reflect that some trees are removed due to a better score occurring with fewer trees.
- [[7140]](https://github.com/h2oai/h2o-3/issues/7140) - Fixed `fit_params` passthrough for scikit-learn compatibility.
- [[7718]](https://github.com/h2oai/h2o-3/issues/7718) - Fixed validateWithCheckpoint to work with default parameter settings.
- [[7719]](https://github.com/h2oai/h2o-3/issues/7719) - Fixed validateWithCheckpoint to work with parameters that are arrays.
Improvement
- [[6990]](https://github.com/h2oai/h2o-3/issues/6990) - Added expert option to force-enable MOJO for CoxPH even when `interactions` are enabled.
- [[7060]](https://github.com/h2oai/h2o-3/issues/7060) - Makes language rules generation on demand and introduced “EnumLimited” option for categorical encoding.
New Feature
- [[6968]](https://github.com/h2oai/h2o-3/issues/6968) - Added `transform_frame` for GLRM allowing users to obtain the new X for a new data set.
- [[6974]](https://github.com/h2oai/h2o-3/issues/6974) - Added support for numerical interactions in CoxPH MOJO.
Docs
- [[6965]](https://github.com/h2oai/h2o-3/issues/6965) - Fixed the `uplift_metric` documentation for Uplift DRF.
- [[6967]](https://github.com/h2oai/h2o-3/issues/6967) - Added `transform_frame` to GLRM documentation.
- [[6970]](https://github.com/h2oai/h2o-3/issues/6970) - Added `mode = “maxrsweep”` to ModelSelection documentation.
- [[6971]](https://github.com/h2oai/h2o-3/issues/6971) - Corrected the R documentation on R^2.
- [[6988]](https://github.com/h2oai/h2o-3/issues/6988) - Updated supported MOJO list to include GAM MOJO import.
Security
- [[6978]](https://github.com/h2oai/h2o-3/issues/6978) - Fixed security issue in genmodel (CVE-2022-25647).
Zumbo (3.36.1.2) - 5/26/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/2/index.html</a>
Bug
- [[6999]](https://github.com/h2oai/h2o-3/issues/6999) - Refactored Uplift DRF methods.
- [[7001]](https://github.com/h2oai/h2o-3/issues/7001) - Removed duplicate runs in MaxR.
- [[7014]](https://github.com/h2oai/h2o-3/issues/7014) - Fixed the ambiguity check in Explain’s consolidate varimp.
- [[7020]](https://github.com/h2oai/h2o-3/issues/7020) - Improved efficiency of ``pd_plot`` and ``ice_plot`` and made ``rug`` optional.
- [[7021]](https://github.com/h2oai/h2o-3/issues/7021) - Fixed H2O failing with null pointer exception when providing an improper ``-network`` to h2o.jar.
- [[7022]](https://github.com/h2oai/h2o-3/issues/7022) - Fixed external XGBoost on K8s.
- [[7042]](https://github.com/h2oai/h2o-3/issues/7042) - Fixed failing concurrent-running GLM training processes.
- [[7283]](https://github.com/h2oai/h2o-3/issues/7283) - Fixed missing time values SHAP summary plot error.
- [[7732]](https://github.com/h2oai/h2o-3/issues/7732) - Fixed Partial Dependence Plot’s date/time handling from explainability modules.
New Feature
- [[7004]](https://github.com/h2oai/h2o-3/issues/7004) - Updated Uplift DRF API.
Docs
- [[7000]](https://github.com/h2oai/h2o-3/issues/7000) - Added ``model_summary`` examples for GLM.
- [[7010]](https://github.com/h2oai/h2o-3/issues/7010) - Updated incorrect formula in GLM booklet.
- [[7219]](https://github.com/h2oai/h2o-3/issues/7219) - Updated Python Module documentation readability.
Zumbo (3.36.1.1) - 4/13/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zumbo/1/index.html</a>
Bug
- [[7035]](https://github.com/h2oai/h2o-3/issues/7035) - Fixed Residual Analysis plot flipping the residual calculation.
- [[7040]](https://github.com/h2oai/h2o-3/issues/7040) - Added more detailed exception when disconnected due to error caused by `Rcurl`.
- [[7047]](https://github.com/h2oai/h2o-3/issues/7047) - Made R client attempt to connect to `curl` package instead of `Rcurl` package first.
- [[7055]](https://github.com/h2oai/h2o-3/issues/7055) - Ensures GLM models fail instead of throwing warnings when `beta_contraints` and `non_negative are used with multinomial or ordinal families.
- [[7057]](https://github.com/h2oai/h2o-3/issues/7057) - Fixed how `cv_computeAndSetOptimalParameters` deals with multiple `alpha` and `lambda` values across different folds.
- [[7058]](https://github.com/h2oai/h2o-3/issues/7058) - Increased MaxR running speed.
- [[7068]](https://github.com/h2oai/h2o-3/issues/7068) - Fixed `getGLMRegularizationPath` erroring out when `standardize = False`.
- [[7194]](https://github.com/h2oai/h2o-3/issues/7194) - Fixed Keystore not generating on Java 16+.
- [[7634]](https://github.com/h2oai/h2o-3/issues/7634) - Added a `num_of_features` argument to `h2o.varimp_heatmap` to limit the number of displayed variables.
- [[12130]](https://github.com/h2oai/h2o-3/issues/12130) - Fixed `cross_validation_metrics_summary` not being accessible for Stacked Ensemble.
Improvement
- [[7029]](https://github.com/h2oai/h2o-3/issues/7029) - Improved AUUC result information in Uplift DRF by adding information on number of bins.
- [[7038]](https://github.com/h2oai/h2o-3/issues/7038) - Replaced `class()` with `inherits()` in R package.
- [[7039]](https://github.com/h2oai/h2o-3/issues/7039) - Fixed invalid URLs in R Package.
- [[7051]](https://github.com/h2oai/h2o-3/issues/7051) - Added normalized `AUUC` to Uplift DRF.
- [[7059]](https://github.com/h2oai/h2o-3/issues/7059) - Sped-up AutoML by avoiding sleep-waiting.
- [[7061]](https://github.com/h2oai/h2o-3/issues/7061) - Removed Stacked Ensembles with XGB metalearner to increase speed.
- [[7062]](https://github.com/h2oai/h2o-3/issues/7062) - Ensures AutoML reproducibility when `max_models` is used.
- [[7135]](https://github.com/h2oai/h2o-3/issues/7135) - Updated AutoML default leaderboard regression sorting to `RMSE`.
New Feature
- [[7036]](https://github.com/h2oai/h2o-3/issues/7036) - Bundled several basic datasets with H2O for use in examples.
- [[7050]](https://github.com/h2oai/h2o-3/issues/7050) - Added h2o.jar assembly for secure Steam deployments and excluded PAM authentication from minimal/Steam builds.
- [[7054]](https://github.com/h2oai/h2o-3/issues/7054) - Added ability to ingest data from secured Hive using h2odriver.jar in standalone.
- [[7073]](https://github.com/h2oai/h2o-3/issues/7073) - Bundled KrbStandalone extension in h2odriver.jar.
- [[7085]](https://github.com/h2oai/h2o-3/issues/7085) - Implemented new method for defining histogram split-points in GBM/DRF designed to address outlier issues with default `UniformAdaptive` method.
- [[7092]](https://github.com/h2oai/h2o-3/issues/7092) - Added ability to reorder frame levels based on their frequencies and to relieve only topN levels for GLM.
- [[7094]](https://github.com/h2oai/h2o-3/issues/7094) - Added a function to calculate predicted versus actual response in GLM.
- [[7258]](https://github.com/h2oai/h2o-3/issues/7258) - Implemented MOJO for Extended Isolation Forest.
- [[7261]](https://github.com/h2oai/h2o-3/issues/7261) - Added monotone splines to GAM.
- [[7271]](https://github.com/h2oai/h2o-3/issues/7271) - Added a plot function for gains/lift to R and Python.
- [[7285]](https://github.com/h2oai/h2o-3/issues/7285) - Added ability to acquire metric builder updates for Sparkling Water calculation without H2O runtime.
- [[7664]](https://github.com/h2oai/h2o-3/issues/7664) - Added support for `interaction_constraints` to GBM.
- [[7785]](https://github.com/h2oai/h2o-3/issues/7785) - Exposed distribution parameter in AutoML
Task
- [[7078]](https://github.com/h2oai/h2o-3/issues/7078) - Decoupled Infogram and XGBoost removing Infograms reliance on XGBoost to work.
- [[7080]](https://github.com/h2oai/h2o-3/issues/7080) - Verified GLM binomial IRLSM implementation and `p-value` calculation.
- [[7089]](https://github.com/h2oai/h2o-3/issues/7089) - Added private ModelBuilder parameter to AutoML to enforce the time budget on the final model after cross-validation.
Sub-Task
- [[7069]](https://github.com/h2oai/h2o-3/issues/7069) - Made Ice plot functionalities also available on `pd_plot`.
- [[7160]](https://github.com/h2oai/h2o-3/issues/7160) - Added option to normalize y-axis values.
- [[7163]](https://github.com/h2oai/h2o-3/issues/7163) - Added option to display logodds for binary models for Ice plots.
- [[7164]](https://github.com/h2oai/h2o-3/issues/7164) - Added ability to save final graphing data to a frame for Ice plots.
- [[7165]](https://github.com/h2oai/h2o-3/issues/7165) - Added option to specify a grouping variable for Ice plots.
- [[7166]](https://github.com/h2oai/h2o-3/issues/7166) - Shows original observation values as points on the line for Ice plots.
- [[7167]](https://github.com/h2oai/h2o-3/issues/7167) - Added option to toggle PDP vs Ice lines on or off.
Docs
- [[7034]](https://github.com/h2oai/h2o-3/issues/7034) - Added documentation on the monotone spline for GAM.
- [[7027]](https://github.com/h2oai/h2o-3/issues/7027) - Added links to the Additional Resources page to the sites where users can ask questions.
- [[7031]](https://github.com/h2oai/h2o-3/issues/7031) - Updated the examples for the Residual Analysis Plot.
- [[7037]](https://github.com/h2oai/h2o-3/issues/7037) - Updated the K8s deployment tutorial.
- [[7043]](https://github.com/h2oai/h2o-3/issues/7043) - Improved Uplift DRF User Guide documentation.
- [[7049]](https://github.com/h2oai/h2o-3/issues/7049) - Shifted the links from the H2O-3 docs page to the User Guide “Additional Resources” page.
- [[7066]](https://github.com/h2oai/h2o-3/issues/7066) - Fixed MOJO importable/exportable table in User Guide.
- [[7067]](https://github.com/h2oai/h2o-3/issues/7067) - Added a note that MOJOs won’t build if `interactions` are specified.
- [[7070]](https://github.com/h2oai/h2o-3/issues/7070) - Added information on how H2O handles date columns.
- [[7075]](https://github.com/h2oai/h2o-3/issues/7075) - Fixed code typos on Admissible ML page in User Guide.
- [[7074]](https://github.com/h2oai/h2o-3/issues/7074) - Added information on the `-hdfs_config` tag.
Zorn (3.36.0.4) - 3/30/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/4/index.html</a>
Bug
- [[7046]](https://github.com/h2oai/h2o-3/issues/7046) - Fixed logic operations error in R package.
- [[7052]](https://github.com/h2oai/h2o-3/issues/7052) - Clarified that `enum` and `eigen` `categorical_encoding` values do not work for XGBoost.
Improvement
- [[7177]](https://github.com/h2oai/h2o-3/issues/7177) - Added the Qini value metric to Uplift DRF.
Docs
- [[7157]](https://github.com/h2oai/h2o-3/issues/7157) - Added information on the `make_metrics` command to the Performance and Prediction section of the User Guide.
Zorn (3.36.0.3) - 2/16/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/3/index.html</a>
Bug
- [[7098]](https://github.com/h2oai/h2o-3/issues/7098) - Fixed S3 file downloads not working by adding `aws_java_sdk_sts` as a dependency of H2O persist S3.
- [[7102]](https://github.com/h2oai/h2o-3/issues/7102) - Added note to GBM, DRF, IF, and EIF that `build_tree_one_node=True` does not work with current release.
- [[7108]](https://github.com/h2oai/h2o-3/issues/7108) - Extended AWS default credential chain instead of replacing it.
- [[7112]](https://github.com/h2oai/h2o-3/issues/7112) - Fixed import failures for URLs longer than 152 characters.
- [[7113]](https://github.com/h2oai/h2o-3/issues/7113) - Fix AutoML ignoring `verbosity` setting.
- [[7138]](https://github.com/h2oai/h2o-3/issues/7138) - Fixed Huber distribution bug for `deviance`.
Improvement
- [[7187]](https://github.com/h2oai/h2o-3/issues/7187) - Removed “H2O API Extensions” from `h2o.init()` output.
Docs
- [[7096]](https://github.com/h2oai/h2o-3/issues/7096) - Corrected typos and inconsistencies in Admissible ML documentation.
- [[7119]](https://github.com/h2oai/h2o-3/issues/7119) - Updated copyright year in documentation.
- [[7128]](https://github.com/h2oai/h2o-3/issues/7128) - Clarified feasible intervals for tweedie power.
- [[7196]](https://github.com/h2oai/h2o-3/issues/7196) - Clarified Java requirements when running H2O on Hadoop.
Zorn (3.36.0.2) - 1/25/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/2/index.html</a>
Bug
- [[7125]](https://github.com/h2oai/h2o-3/issues/7125) - Updated XGBoostMojoModel to only consider the number of built trees, not the value of ``ntrees``.
- [[7126]](https://github.com/h2oai/h2o-3/issues/7126) - Fixed issue in AutoEncoder’s early stopping automatic selection by setting ``AUTO = MSE`` instead of ``deviance``.
- [[7133]](https://github.com/h2oai/h2o-3/issues/7133) - Fixed MOJO imports to retain information on weights column.
- [[7143]](https://github.com/h2oai/h2o-3/issues/7143) - Fixed XGBoost errors on Infogram by improving support for XGBoost.
- [[7145]](https://github.com/h2oai/h2o-3/issues/7145) - Fixed MOJO import automatically re-using original Model ID for current release cycle.
- [[7149]](https://github.com/h2oai/h2o-3/issues/7149) - Fixed import of Parquet files from S3.
- [[7155]](https://github.com/h2oai/h2o-3/issues/7155) - Fixed `h2o.group_by` warning present in documentation example caused by function only reading the first column when several are provided.
- [[7174]](https://github.com/h2oai/h2o-3/issues/7174) - Added check to ensure that a model supports MOJOs to prevent production of bad MOJOs.
- [[7179]](https://github.com/h2oai/h2o-3/issues/7179) - Fixed Python warnings before model training when training with offset, weights, and fold columns.
- [[7181]](https://github.com/h2oai/h2o-3/issues/7181) - Fixed MOJO upload in Python.
- [[7201]](https://github.com/h2oai/h2o-3/issues/7201) - Fixed error in uploading pandas DataFrame to H2O by enforcing `uft-8` encoding.
- [[7273]](https://github.com/h2oai/h2o-3/issues/7273) - Customized FormAuthenticator to use relative redirects.
Improvement
- [[7141]](https://github.com/h2oai/h2o-3/issues/7141) - Removed numpy dependency for Infogram.
New Feature
- [[7232]](https://github.com/h2oai/h2o-3/issues/7232) - Added backward selection method for ModelSelection.
Task
- [[7123]](https://github.com/h2oai/h2o-3/issues/7123) - Added support to PredictCsv for testing concurrent predictions.
Docs
- [[7136]](https://github.com/h2oai/h2o-3/issues/7136) - Added backward mode documentation to ModelSelection.
- [[7194]](https://github.com/h2oai/h2o-3/issues/7194) - Updated Kubernetes Headless Service and StatefulSet documentation.
Zorn (3.36.0.1) - 12/29/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zorn/1/index.html</a>
Bug
- [[7214]](https://github.com/h2oai/h2o-3/issues/7214) - Fixed differences in H2O’s random behavior across Java versions by disabling Stream API in this task.
- [[7247]](https://github.com/h2oai/h2o-3/issues/7247) - Fixed CoxPH summary method in Python to return H2OTwoDimTable.
- [[7273]](https://github.com/h2oai/h2o-3/issues/7273) - Fixed form authentication not working by enforcing relative redirects in Jetty.
- [[7888]](https://github.com/h2oai/h2o-3/issues/7888) - Fixed exception raised in K-Means when a model is built using `nfolds` by disabling centroid stats for Cross-Validation.
Improvement
- [[7188]](https://github.com/h2oai/h2o-3/issues/7188) - Removed `ymu` and `rank` visibility from FlowUI.
- [[7209]](https://github.com/h2oai/h2o-3/issues/7209) - Exposed `lambda` in Rulefit to have better control over regularization strength.
- [[7217]](https://github.com/h2oai/h2o-3/issues/7217) - Implemented sequential replacement method with ModelSelection.
- [[7222]](https://github.com/h2oai/h2o-3/issues/7222) - Improved rule extraction from trees in RuleFit.
- [[7240]](https://github.com/h2oai/h2o-3/issues/7240) - Improved exception handling in AutoML and Grids to prevent model failure.
- [[7395]](https://github.com/h2oai/h2o-3/issues/7395) - Ensured Infogram uses validation frame and cross-validation when enabled.
- [[8096]](https://github.com/h2oai/h2o-3/issues/8096) - Added dynamic stacking metalearning strategy for Stacked Ensemble in AutoML.
New Feature
- [[7246]](https://github.com/h2oai/h2o-3/issues/7246) - Added support and rule coverage to RuleFit.
- [[7268]](https://github.com/h2oai/h2o-3/issues/7268) - Added support for importing GAM MOJO.
- [[7279]](https://github.com/h2oai/h2o-3/issues/7279) - Added a convenience tool that converts MOJO to POJO from the command line.
- [[7280]](https://github.com/h2oai/h2o-3/issues/7280) - Added support allowing users to modify floating point representation in POJO.
- [[7287]](https://github.com/h2oai/h2o-3/issues/7287) - Added experimental support for importing POJO for in-H2O scoring.
- [[7316]](https://github.com/h2oai/h2o-3/issues/7316) - Added official support for Java 16 and 17.
- [[7323]](https://github.com/h2oai/h2o-3/issues/7323) - Added Java 16 and 17 to the cluster.
- [[7333]](https://github.com/h2oai/h2o-3/issues/7333) - Added a compatibility K8s module that allows older versions of H2O to run on K8s.
- [[7447]](https://github.com/h2oai/h2o-3/issues/7447) - Added ability to convert MOJO to POJO for tree models.
- [[7515]](https://github.com/h2oai/h2o-3/issues/7515) - Added support enabling users to configure S3 with S3A configuration.
- [[7574]](https://github.com/h2oai/h2o-3/issues/7574) - Implemented the Infogram model.
- [[11818]](https://github.com/h2oai/h2o-3/issues/11818) - Implemented the Uplift DRF algorithm.
Task
- [[7322]](https://github.com/h2oai/h2o-3/issues/7322) - Upgraded to Gradle 7 to support Java 16+.
- [[7430]](https://github.com/h2oai/h2o-3/issues/7430) - Added R API for Infogram.
Docs
- [[7212]](https://github.com/h2oai/h2o-3/issues/7212) - Added documentation on Infogram to the User Guide.
- [[7279]](https://github.com/h2oai/h2o-3/issues/7279) - Added documentation on ModelSelection to the User Guide.
- [[7275]](https://github.com/h2oai/h2o-3/issues/7275) - Added notebook on floating point issue for POJO and FAQ documentation on POJO split points.
- [[7329]](https://github.com/h2oai/h2o-3/issues/7329) - Fixed bullet list formatting issues.
- [[7742]](https://github.com/h2oai/h2o-3/issues/7742) - Updated R Reference Guide list.
Zizler (3.34.0.8) - 1/13/2022
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/8/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/8/index.html</a>
Bug
- [[7148]](https://github.com/h2oai/h2o-3/issues/7148) - Fixed MOJO import automatically re-using original Model ID.
Security
- [[7147]](https://github.com/h2oai/h2o-3/issues/7147) - Upgraded to log4j 2.17.1.
Zizler (3.34.0.7) - 12/21/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/7/index.html</a>
Security
- Fixed CVE-2021-45105 log4j vulnerability.
Zizler (3.34.0.6) - 12/15/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/6/index.html</a>
Security
- Fixed CVE-2021-45046 log4j vulnerability.
Zizler (3.34.0.5) - 12/13/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/5/index.html</a>
Bug
- [[7213]](https://github.com/h2oai/h2o-3/issues/7213) - Fixed permutation variable importance to correctly work with weights.
- [[7234]](https://github.com/h2oai/h2o-3/issues/7234) - Fixed data removal issue in GAM caused by fitting two different models on the same DataFrame.
Improvement
- [[7233]](https://github.com/h2oai/h2o-3/issues/7233) - Added `coef()` and `coef_norm()` functions to MaxRGLM.
- [[7251]](https://github.com/h2oai/h2o-3/issues/7251) - Added ability that labels observations that match rules in Rulefit.
- [[7262]](https://github.com/h2oai/h2o-3/issues/7262) - Updated parquet parser to handle dates allowing H2O `import_file()` to import date columns from Spark DataFrame.
- [[7276]](https://github.com/h2oai/h2o-3/issues/7276) - Consolidated Rulefit rules to remove unnecessary splits.
- [[7439]](https://github.com/h2oai/h2o-3/issues/7439) - Improved the efficiency of job polling in AutoML.
- [[7474]](https://github.com/h2oai/h2o-3/issues/7474) - Deduplicated Rulefit rules in post-processing step.
New Feature
- [[7235]](https://github.com/h2oai/h2o-3/issues/7235) - Added option to mimic the “ActiveProcessorCount” for older JVMs.
Task
- [[7227]](https://github.com/h2oai/h2o-3/issues/7227) - Added warning in GLRM for when users set `model_id` and `representation_name` to the same string to help avoid a collision of model and frame using the same key.
- [[7239]](https://github.com/h2oai/h2o-3/issues/7239) - Added `rank` and `ymu` model outputs to GLM.
Docs
- [[7210]](https://github.com/h2oai/h2o-3/issues/7210) - Added link to the Change Log in the User Guide index.
- [[7218]](https://github.com/h2oai/h2o-3/issues/7218) - Updated parameter list for MaxRGLM and outlined that MaxRGLM only support regression.
- [[7228]](https://github.com/h2oai/h2o-3/issues/7228) - Updated MaxRGLM examples to use new functions `coef()`, `coef_norm()`, and `result()`.
- [[7236]](https://github.com/h2oai/h2o-3/issues/7236) - Added examples in R/Python on how to get reproducibility information.
- [[7296]](https://github.com/h2oai/h2o-3/issues/7296) - Fixed local build warnings for Python Module documentation.
Security
- Upgraded to log4j 2.15.0 to address vulnerability CVE-2021-44228.
Zizler (3.34.0.4) - 11/17/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/4/index.html</a>
Bug
- [[7252]](https://github.com/h2oai/h2o-3/issues/7252) - Fixed broken `weights_column` in GAM.
- [[7256]](https://github.com/h2oai/h2o-3/issues/7256) - Fixed printing a DRF model when there are no out-of-bag samples.
- [[7259]](https://github.com/h2oai/h2o-3/issues/7259) - Fixed the `pyunit_PUBDEV_5008_5386_glm_ordinal_large.py` test from failing.
- [[7263]](https://github.com/h2oai/h2o-3/issues/7263) - Fixed AutoML XGBoost `learn_rate` search step.
- [[7265]](https://github.com/h2oai/h2o-3/issues/7265) - Ensured that jobs are rendered correctly in Flow and that AutoML internal jobs can be monitored without crashing on the backend.
- [[7269]](https://github.com/h2oai/h2o-3/issues/7269) - Fixed `gam_columns` failure in the `pyunit_PUBDEV_7185_GAM_mojo_ordinal.py` test.
- [[7290]](https://github.com/h2oai/h2o-3/issues/7290) - Outlined that `tree_method=“approx”` is not supported with `col_sample_rate` or `col_sample_by_level` in XGBoost.
- [[7517]](https://github.com/h2oai/h2o-3/issues/7517) - Fixed multinomial classification in Rulefit.
- [[7681]](https://github.com/h2oai/h2o-3/issues/7681) - Fixed inconsistencies in GLM `beta_constraints`.
- [[7738]](https://github.com/h2oai/h2o-3/issues/7738) - Enabled ability to provide metalearner parameters for NaiveBayes and XGBoost.
Improvement
- [[7358]](https://github.com/h2oai/h2o-3/issues/7358) - Added a custom model ID parameter to MOJO importing/uploading methods through R/Python API and if a custom model ID is not specified, the default model ID is propagated as the models name from the MOJO path.
- [[7361]](https://github.com/h2oai/h2o-3/issues/7361) - Added warning for users who accidentally build a regression model when attempting building a binary classification model because they forgot to convert their target to categorical.
- [[7372]](https://github.com/h2oai/h2o-3/issues/7372) - Tuned `scale_pos_weight` parameter for XGBooost in AutoML for imbalanced data.
New Feature
- [[7274]](https://github.com/h2oai/h2o-3/issues/7274) - Added saving parameters to plot functions.
Task
- [[7250]](https://github.com/h2oai/h2o-3/issues/7250) - Added GAM training/validation metrics.
- [[7264]](https://github.com/h2oai/h2o-3/issues/7264) - Ensured H2O-3 builds with pip version >= 21.3.
- [[7311]](https://github.com/h2oai/h2o-3/issues/7311) - Added result frame to MAXRGLM.
Docs
- [[7267]](https://github.com/h2oai/h2o-3/issues/7267) - Localized MOJO support list for all the H2O-3 algorithms.
- [[7278]](https://github.com/h2oai/h2o-3/issues/7278) - Added Gains/Lift documentation to the Performance and Prediction section of the User Guide.
- [[7288]](https://github.com/h2oai/h2o-3/issues/7288) - Corrected metric in the Performance and Prediction “Sensitive to Outliers” section of the User Guide.
- [[7377]](https://github.com/h2oai/h2o-3/issues/7377) - Clarified that `asnumeric()` converted ‘enum’ columns to underlying factor values and highlighted correct transformation approach.
Zizler (3.34.0.3) - 10/7/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/3/index.html</a>
Bug
- [[7291]](https://github.com/h2oai/h2o-3/issues/7291) - Fixed user login from key tab in standalone on Kerberos.
- [[7300]](https://github.com/h2oai/h2o-3/issues/7300) - Improved error messages in Explain module by making the errors clearer.
- [[7307]](https://github.com/h2oai/h2o-3/issues/7307) - Fixed H2OTable colTypes in Grid’s summary table.
- [[7308]](https://github.com/h2oai/h2o-3/issues/7308) - Fixed infinite loop in hex.grid.HyperSpaceWalker.RandomDiscreteValueWalker.
- [[7317]](https://github.com/h2oai/h2o-3/issues/7317) - Fixed AutoML ignoring optional Stacked Ensembles.
- [[7319]](https://github.com/h2oai/h2o-3/issues/7319) - Fixed NPE thrown in AutoML when XGBoost is disabled/not available.
- [[7320]](https://github.com/h2oai/h2o-3/issues/7320) - Fixed CRAN install.
- [[8458]](https://github.com/h2oai/h2o-3/issues/8458) - Improved XGBoost API to ensure both `col_sample_rate` and `colsample_bylevel` (and other XGBoost parameters aliases) are set correctly.
- [[7375]](https://github.com/h2oai/h2o-3/issues/7375) - Fixed NPE thrown for `ModelJsonReader.findINJson` for cases when path does not exist.
Improvement
- [[7314]](https://github.com/h2oai/h2o-3/issues/7314) - Exposed AutoML `get_leaderboard` as a method in Python.
- [[7315]](https://github.com/h2oai/h2o-3/issues/7315) - Improved Python client by printing the stacktrace in case of ServerError allowing users to report informative issues for investigation.
- [[7350]](https://github.com/h2oai/h2o-3/issues/7350) - Enhanced tests by testing the case through all encodings.
Task
- [[7312]](https://github.com/h2oai/h2o-3/issues/7312) - Updated ANOVA GLM to save model summary as a frame.
- [[7326]](https://github.com/h2oai/h2o-3/issues/7326) - Added GLM offset column support to GLM MOJO.
Docs
- [[7302]](https://github.com/h2oai/h2o-3/issues/7302) - Updated the R/Python AutoML documentation parameters to match the descriptions in the User Guide.
- [[7304]](https://github.com/h2oai/h2o-3/issues/7304) - Removed GLM from `balance_classes` parameter appendix page in the User Guide.
- [[7309]](https://github.com/h2oai/h2o-3/issues/7309) - Updated the `asfactor` procedure documentation to show multiple column usage.
Zizler (3.34.0.1) - 9/14/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zizler/1/index.html</a>
Bug
- [[7330]](https://github.com/h2oai/h2o-3/issues/7330) - Fixed matplotlib 3.4 compatibility issues with `partial_plot`.
- [[7339]](https://github.com/h2oai/h2o-3/issues/7339) - Deprecated `is_supervised` parameter for h2o.grid method in R.
- [[7341]](https://github.com/h2oai/h2o-3/issues/7341) - Fixed AutoML NPE by ensuring that models without metrics are not added to the leaderboard.
- [[7360]](https://github.com/h2oai/h2o-3/issues/7360) - Redistributed the time budget for AutoML.
- [[7365]](https://github.com/h2oai/h2o-3/issues/7365) - Fixed and reorganized the H2O Explain leaderboard and fixed the confusion matrix.
- [[7366]](https://github.com/h2oai/h2o-3/issues/7366) - Decreased the number of displayed features in the heatmap for AutoML inside H2O Explain.
- [[7378]](https://github.com/h2oai/h2o-3/issues/7378) - Fixed NPE raised from ``weight_column`` not being in the training model.
- [[7380]](https://github.com/h2oai/h2o-3/issues/7380) - Fixed the ``weight=0`` documentation change error.
- [[7383]](https://github.com/h2oai/h2o-3/issues/7383) - Fixed failing rotterdam tests.
- [[7387]](https://github.com/h2oai/h2o-3/issues/7387) - Fixed GAM NPE from multiple runs with knots specified in a frame.
- [[8458]](https://github.com/h2oai/h2o-3/issues/8458) - Fixed ``col_sample_rate`` not sampling for XGBoost when set to a value lower than 1.0.
- [[7396]](https://github.com/h2oai/h2o-3/issues/7396) - Fixed wrong column type on MOJO models for Cross-Validation Metrics Summary.
- [[7408]](https://github.com/h2oai/h2o-3/issues/7408) - Prevented R connect from starting H2O locally.
- [[7420]](https://github.com/h2oai/h2o-3/issues/7420) - Added StackedEnsembles to AutoML’s time budget to prevent unexpected training times.
- [[7441]](https://github.com/h2oai/h2o-3/issues/7441) - Fixed the failing `pyunit_scale_pca_rf.py` test.
- [[7475]](https://github.com/h2oai/h2o-3/issues/7475) - Improved AutoML behavior when multiple instances are created in parallel.
- [[7787]](https://github.com/h2oai/h2o-3/issues/7787) - Solved corner cases involving mapping between encoded varimps and predictor columns for H2O Explain by making the varimp feature consolidation more robust.
Improvement
- [[7381]](https://github.com/h2oai/h2o-3/issues/7381) - Ensured that AutoML uses the entire time budget for `max_runtime`.
- [[7455]](https://github.com/h2oai/h2o-3/issues/7455) - Implemented custom progress widgets for Wave apps using H2O-3.
- [[7461]](https://github.com/h2oai/h2o-3/issues/7461) - Allowed users to convert floats to doubles with PrintMojo to prevent possible parsing issues.
- [[7465]](https://github.com/h2oai/h2o-3/issues/7465) - Updated GBM cross validation with ``early_stopping`` to use ``ntrees`` that produce the best score.
- [[7466]](https://github.com/h2oai/h2o-3/issues/7466) - Enabled ``print_mojo`` to produce .png outputs.
- [[7470]](https://github.com/h2oai/h2o-3/issues/7470) - Updated Python API for all algorithms and AutoML to retrieve the trained model or leader.
- [[7476]](https://github.com/h2oai/h2o-3/issues/7476) - Removed algorithm-specific logic from base classes.
- [[7478]](https://github.com/h2oai/h2o-3/issues/7478) - Added support for scoreContributions for imported MOJOs in Java.
- [[7480]](https://github.com/h2oai/h2o-3/issues/7480) - Exposed AutoML args as writeable properties until first called to train.
- [[7482]](https://github.com/h2oai/h2o-3/issues/7482) - Updated XGBoost ``print_mojo`` to now output weights.
- [[7498]](https://github.com/h2oai/h2o-3/issues/7498) - Removed the Python client dependency on colorama.
- [[7504]](https://github.com/h2oai/h2o-3/issues/7504) - Added the parameters and their default values to the ``_init_`` function of the Py code generator.
- [[7535]](https://github.com/h2oai/h2o-3/issues/7535) - Reduced the workspace of the validation frame in GBM by sharing it with the training frame in cross validation.
- [[7564]](https://github.com/h2oai/h2o-3/issues/7564) - Slightly reduced precision of predictions stored in holdout frames to significantly save on memory.
- [[7633]](https://github.com/h2oai/h2o-3/issues/7633) - Removed warning in the Stacked Ensemble prediction function about missing ``fold_column`` frame.
- [[7690]](https://github.com/h2oai/h2o-3/issues/7690) - Enabled returning data from Explain’s `varimp_heatmap` and `model_correlation_matrix`.
- [[7708]](https://github.com/h2oai/h2o-3/issues/7708) - Exposed the ``top n`` and ``bottom n`` reason codes in Python/R and MOJO.
- [[12171]](https://github.com/h2oai/h2o-3/issues/12171) - Fixed nightly build version mismatch that prevented the H2OCluster timezone being set to America/Denver.
New Feature
- [[7336]](https://github.com/h2oai/h2o-3/issues/7336) - Implemented a java-self-check to allow users to run on latest Java.
- [[7343]](https://github.com/h2oai/h2o-3/issues/7343) - Sped up GBM by optimizing the building of histograms.
- [[7368]](https://github.com/h2oai/h2o-3/issues/7368) - Added a warning to the TreeSHAP reweighting feature if there are 0 weights and updated the API.
- [[7418]](https://github.com/h2oai/h2o-3/issues/7418) - Added Maximum R Square Improvement (MAXR) algorithm to GLM.
- [[7424]](https://github.com/h2oai/h2o-3/issues/7424) - Added warning for when H2O doesn’t have enough memory to run XGBoost.
- [[7431]](https://github.com/h2oai/h2o-3/issues/7431) - Added the ability to specify a custom file name when saving a MOJO.
- [[7448]](https://github.com/h2oai/h2o-3/issues/7448) - Added output version number of genmodel.jar when printing usage for PrintMojo.
- [[7536]](https://github.com/h2oai/h2o-3/issues/7536) - Added MOJO to Rulefit.
- [[7550]](https://github.com/h2oai/h2o-3/issues/7550) - Implemented ability to calculate Shapley values on a re-weighted tree.
- [[7561]](https://github.com/h2oai/h2o-3/issues/7561) - Implemented H2O ANOVA GLM algorithm for GLM.
- [[8283]](https://github.com/h2oai/h2o-3/issues/8283) - Improved and consolidated the handling of version mismatch between Python and Backend.
- [[8500]](https://github.com/h2oai/h2o-3/issues/8500) - Implemented permutation feature importance for black-box models.
- [[8501]](https://github.com/h2oai/h2o-3/issues/8501) - Implemented Extended Isolation Forest algorithm.
- [[9260]](https://github.com/h2oai/h2o-3/issues/9260) - Added support for saving a model directly to S3.
Task
- [[7363]](https://github.com/h2oai/h2o-3/issues/7363) - Fixed the time limits for the Merge/Sort benchmark.
- [[7454]](https://github.com/h2oai/h2o-3/issues/7454) - Switched removed pandas ``as_matrix`` method to ``.values`` and exposed the interim `pandas.DataFrame` object.
- [[7533]](https://github.com/h2oai/h2o-3/issues/7533) - Fixed S3 credential for `pyunit_s3_model_save.py` test.
- [[7565]](https://github.com/h2oai/h2o-3/issues/7565) - Connected XGBoost aggregation functionality with sorting functionality.
Technical task
- [[7449]](https://github.com/h2oai/h2o-3/issues/7449) - Replaced subsampling in Extended Isolation Forest.
Docs
- [[7348]](https://github.com/h2oai/h2o-3/issues/7348) - Updated the AutoML FAQ.
- [[7351]](https://github.com/h2oai/h2o-3/issues/7351) - Corrected the ``ignored_columns`` example.
- [[7356]](https://github.com/h2oai/h2o-3/issues/7356) - Added RMarkdown, Jupyter Notebook, and HTML output example files to H2O Explain documentation.
- [[7373]](https://github.com/h2oai/h2o-3/issues/7373) - Added Maximum R Improvements (MAXR) GLM documentation.
- [[7392]](https://github.com/h2oai/h2o-3/issues/7392) - Added the loss function equations for each distribution and link type.
- [[7405]](https://github.com/h2oai/h2o-3/issues/7405) - Updated the documentation about StackedEnsembles time constraints in AutoML.
- [[7446]](https://github.com/h2oai/h2o-3/issues/7446) - Clarified that the Explain function only works for supervised models.
- [[7471]](https://github.com/h2oai/h2o-3/issues/7471) - Added Examine Models section to AutoML documentation.
- [[7484]](https://github.com/h2oai/h2o-3/issues/7484) - Added documentation for H2O ANOVA GLM algorithm.
- [[7526]](https://github.com/h2oai/h2o-3/issues/7526) - Fixed the H2O Explain example in the documentation.
- [[7596]](https://github.com/h2oai/h2o-3/issues/7596) - Updated and gathered Java links to a singular place in the User Guide.
Zipf (3.32.1.7) - 8/31/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/7/index.html</a>
Bug
- [[7419]](https://github.com/h2oai/h2o-3/issues/7419) - Fixed predicting issues with imported MOJOs trained with an offset-column.
- [[7406]](https://github.com/h2oai/h2o-3/issues/7406) - Fixed slow tree building by implementing a switch to turn off the generation of plain language rules.
- [[7357]](https://github.com/h2oai/h2o-3/issues/7357) - Fixed potential NPE thrown by setting `_orig_projection_array=[]`.
- [[7346]](https://github.com/h2oai/h2o-3/issues/7346) - Fixed generic model deserialization.
- [[7345]](https://github.com/h2oai/h2o-3/issues/7345) - Fixed predictions for splits NA vs REST with monotone constraints.
New Feature
- [[7362]](https://github.com/h2oai/h2o-3/issues/7362) - H2O Standalone now uses log4j2 as the logger implementation.
Zipf (3.32.1.6) - 8/19/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/6/index.html</a>
Bug
- [[7390]](https://github.com/h2oai/h2o-3/issues/7390) - Fixed the POJO mismatch from MOJO and in-H2O scoring for an unseen categorical value.
- [[7393]](https://github.com/h2oai/h2o-3/issues/7393) - Simplified duplicated XGBoost parameters in Flow.
- [[7414]](https://github.com/h2oai/h2o-3/issues/7414) - Fixed broken data frame conversion behavior.
Improvement
- Added security updates.
New Feature
- [[7371]](https://github.com/h2oai/h2o-3/issues/7371) - Exposed the ``scale_pos_weight`` parameter in XGBoost.
Task
- [[7412]](https://github.com/h2oai/h2o-3/issues/7412) - Clarified the anomaly score formula used for score calculation within Isolation Forest and Extended Isolation Forest.
Docs
- [[7553]](https://github.com/h2oai/h2o-3/issues/7553) - Added a note on memory usage when using XGBoost to User Guide.
Zipf (3.32.1.5) - 8/4/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/5/index.html</a>
Bug
- [[7399]](https://github.com/h2oai/h2o-3/issues/7399) - Modified legacy Dockerfile to add a non-root user.
- [[7400]](https://github.com/h2oai/h2o-3/issues/7400) - Fixed an issue where running `java -jar h2o.jar -version` failed.
- [[7403]](https://github.com/h2oai/h2o-3/issues/7403) - Fixed an issue where monotone constraints in GBM caused issues when reproducing the model.
- [[7407]](https://github.com/h2oai/h2o-3/issues/7407) - Fixed an issue that caused DRF to create incorrect leaf nodes due to rounding errors.
- [[7409]](https://github.com/h2oai/h2o-3/issues/7409) - Fixed an issue that caused CoxPH MOJO import to fail.
- [[7411]](https://github.com/h2oai/h2o-3/issues/7411) - Fixed an issue where categorical splits NAvsREST were not represented correctly.
- [[7413]](https://github.com/h2oai/h2o-3/issues/7413) - Fixed GBM reproducibility for correlated columns with NAs.
- [[7416]](https://github.com/h2oai/h2o-3/issues/7416) - Fixed h2odriver so that it no longer uses invalid GC options.
- [[7423]](https://github.com/h2oai/h2o-3/issues/7423) - Fixed GenericModel predictions for non-AUTO categorical encodings.
- [[7434]](https://github.com/h2oai/h2o-3/issues/7434) - Fixed H2O interaction outcomes.
- [[7460]](https://github.com/h2oai/h2o-3/issues/7460) - When `remove_collinear_columns=True`, fixed an issue where the dimension of gradient and coefficients changed when predictors were removed.
Docs
- [[7415]](https://github.com/h2oai/h2o-3/issues/7415) - Updated changelog format.
Zipf (3.32.1.4) - 7/8/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/4/index.html</a>
Bug
- [[7427]](https://github.com/h2oai/h2o-3/issues/7427) - Fixed h2odriver invalid argument error on Java 11.
- [[7429]](https://github.com/h2oai/h2o-3/issues/7429) - Fixed GLM `GRADIENT_DESCENT_SQERR` Solver validation.
- [[7433]](https://github.com/h2oai/h2o-3/issues/7433) - Upgraded to latest version of Javassist (3.28).
- [[7444]](https://github.com/h2oai/h2o-3/issues/7444) - Fixed H statistic gpu assertion error.
- [[7456]](https://github.com/h2oai/h2o-3/issues/7456) - Fixed predict contributions failure in multi-MOJO environments.
- [[7457]](https://github.com/h2oai/h2o-3/issues/7457) - Fixed bug in ordinal GLM class predictions.
- [[7462]](https://github.com/h2oai/h2o-3/issues/7462) - Fixed Partial Dependent Plot not working with Flow.
- [[7469]](https://github.com/h2oai/h2o-3/issues/7469) - Updated to current Python syntax.
- [[7483]](https://github.com/h2oai/h2o-3/issues/7483) - Fixed bug in ordinal GLM class predictions.
Improvement
- [[7509]](https://github.com/h2oai/h2o-3/issues/7509) - Added support for refreshing HDFS delegation tokens for standalone H2O.
New Feature
- [[7540]](https://github.com/h2oai/h2o-3/issues/7540) - Obtained Friedman’s H statistic for XGBoost and GBM.
Task
- [[7500]](https://github.com/h2oai/h2o-3/issues/7500) - Added a warning message when using `alpha` as a hyperparameter for GLM
Docs
- [[7492]](https://github.com/h2oai/h2o-3/issues/7492) - Added section on how to delete objects in Flow.
- [[7499]](https://github.com/h2oai/h2o-3/issues/7499) - Added a note to the productionizing docs that C++ is only available with additional support.
Zipf (3.32.1.3) - 5/19/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/3/index.html</a>
Bug
- [[7514]](https://github.com/h2oai/h2o-3/issues/7514) - Fixed the printing for `auc_pr` and `pr_auc` in cross-validation summaries.
New Feature
- [[7519]](https://github.com/h2oai/h2o-3/issues/7519) - Added parameter `auc_type` to performance method to compute multiclass AUC.
Task
- [[7503]](https://github.com/h2oai/h2o-3/issues/7503) - Upgraded XGBoost predictor to 0.3.18.
- [[7505]](https://github.com/h2oai/h2o-3/issues/7505) - Increased the timeout duration on the R package jar download.
Docs
- [[7530]](https://github.com/h2oai/h2o-3/issues/7530) - Fixed formatting errors for local builds.
- [[7558]](https://github.com/h2oai/h2o-3/issues/7558) - Updated docs examples for baseline hazard, baseline survival, and concordance.
Zipf (3.32.1.2) - 4/29/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/2/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7800'>#7800</a>] - Stacked Ensemble will no longer ignore a column if any base model uses it.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7791'>#7791</a>] - Added a user-friendly reminder that the new explainability functions require newer versions of `ggplot2` in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7698'>#7698</a>] - NullPointerException error no longer thrown when used a saved and reloaded RuleFit model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7693'>#7693</a>] - Can now extract metrics from the validation dataset with a Rulefit Model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7573'>#7573</a>] - Fixed failures from Stacked Ensemble with Multinomial GLM within tests.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7572'>#7572</a>] - Fixed AutoML error when an alpha array is used for GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7570'>#7570</a>] - Fixed “Rollup not possible" stats failure in GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7552'>#7552</a>] - H2O will now still start despite system properties that begin with ‘ai.h2o.’.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7551'>#7551</a>] - H2O exits without logging any buffered messages instead of throwing a NullPointerException when starting H2O with an invalid argument.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7549'>#7549</a>] - ModelDescriptor field in MOJO is now Serializable.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7547'>#7547</a>] - AutoML no longer crashes if model builder produces H2OIllegalArgumentException in the parameter validation phase.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7543'>#7543</a>] - Weights in GLM grid search is no longer used as features.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7529'>#7529</a>] - Fixed Stacked Ensemble MOJO for cases when sub-model doesn’t have the same columns as the metalearner.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7524'>#7524</a>] - Efron-method now fully deterministic in CoxPH.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7562'>#7562</a>] - User now allowed to specify the escape character for parsing CSVs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7557'>#7557</a>] - Added H2O reconnection script for intermittent 401 errors to R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7548'>#7548</a>] - Added ‘ice_root’ error documented in FAQ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7531'>#7531</a>] - Added further regularization to the GLM metalearner.
</li>
</ul>
New Feature
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9372'>#9372</a>] - Warning now issued against irreproducible model when early stopping is enabled but neither `score_tree_interva`l or `score_each_iteration` are defined.
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7625'>#7625</a>] - Encrypted files that contain CSVs can now be imported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7577'>#7577</a>] - Added guidelines for correct use of `remove_collinear_columns` for GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7538'>#7538</a>] - Support added for CDP 7.2.
</li>
</ul>
Docs
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7582'>#7582</a>] - Added information about the `path` argument for exporting .xlsx files.
</li>
</ul>
Zipf (3.32.1.1) - 3/25/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zipf/1/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9268'>#9268</a>] - GBM histograms now ignore rows with NA responses.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8032'>#8032</a>] - Variable Importances added to GLM Generic model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7859'>#7859</a>] - Fixed the ArrayIndexOutOfBoundsException issue with GLM CV.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7816'>#7816</a>] - CoxPH performance no longer fails when a factor is used for the `event_column`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7801'>#7801</a>] - Existing frame no longer overwritten when data with the same query is loaded.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7736'>#7736</a>] - Fixed how `gain` is calculated in XGBFI for GBM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7711'>#7711</a>] - Improved the error messages for `save_to_hive_table`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7685'>#7685</a>] - Added missing argument ’test’ for `h2o.explain_row()`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7665'>#7665</a>] - All trees now supported for XGBoost Print MOJO in Java.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7657'>#7657</a>] - CoxPH `prediction` no longer fails when `offset_column` is specified.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7678'>#7678</a>] - Added keys for Individual Conditional Expectation (ICE) plot in H2OExplanation class.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7635'>#7635</a>] - `modelmodel$parameters$x` now reports actual feature names instead of `names`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7632'>#7632</a>] - `h2o.explain` no longer errors when AutoML object is trained with a `fold_column`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7603'>#7603</a>] - Fixed issues with python’s explanation plots not displaying fully.
</li>
</ul>
New Feature
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7933'>#7933</a>] - Ignored columns that are actually used for model training are unignored and no longer prevent model training to start in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7904'>#7904</a>] - Added baseline hazard function estimate to CoxPH model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7891'>#7891</a>] - Target Encoding now supports feature interactions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7837'>#7837</a>] - Added CoxPH concordance to both Flow and R/Python CoxPH summaries.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7821'>#7821</a>] - Added a `topbasemodel` attribute to AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7811'>#7811</a>] - Added new learning curve plotting function to R/Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7788'>#7788</a>] - Added script for estimating the memory usage of a dataset.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7784'>#7784</a>] - Added fault protections to grid search allowing saving of data and parameters, model checkpointing, and auto-recovery.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7761'>#7761</a>] - Added support for Java 15.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7673'>#7673</a>] - Added CDP7.1 support.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7666'>#7666</a>] - Added support for XGBoost to Print MOJO as JSON.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7627'>#7627</a>] - Added support for refreshing HDFS delegation tokens.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7613'>#7613</a>] - Reverted XGBoost categorical encodings for contributions.
</li>
</ul>
Task
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8002'>#8002</a>] - `max_hit_ratio_k` deprecated and removed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7751'>#7751</a>] - Added upper bound cap to supported Java version in H2O CRAN package requirements.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8165'>#8165</a>] - Users now allowed to include categorical column name in beta constraints.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8059'>#8059</a>] - Multinomial PDP can now be plotted for more than one target class in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7903'>#7903</a>] - Sped up CoxPH concordance score by using tree instead of the direct approach.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7822'>#7822</a>] - XGBoost no longer fails when specifying custom `fold_column`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7799'>#7799</a>] - XGBoost CV models now built on multiple GPUs in parallel.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8459'>#8459</a>] - Missing metrics added to GLM scoring history.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7631'>#7631</a>] - Added validation checks for sampling rates for XGBoost for the R/Python clients.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7624'>#7624</a>] -
No longer errors when trying to use a fold column where not all folds are represented.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7616'>#7616</a>] - Added the `metalearner_transform` option to Stacked Ensemble.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7592'>#7592</a>] - GBM main model now built in parallel to the CV models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7589'>#7589</a>] - Removed redundant extraction weights from GBM/DRF histogram.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7588'>#7588</a>] - GBM now avoids scoring the last iteration twice when early stopping is enabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7586'>#7586</a>] - POJO predictions for XGBoost now even closer to in-H2O predictions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7585'>#7585</a>] - Double-scoring of CV models in AutoML now avoided thus speeding up AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7579'>#7579</a>] - AutoML now uses fewer neurons in DL grids and has improved the metalearner for Stacked Ensemble.
</li>
</ul>
Technical task
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7783'>#7783</a>] - Thin plate regression splines added to GAM.
</li>
</ul>
Docs
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7728'>#7728</a>] - Added checkpoint description to GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7680'>#7680</a>] - Added thin plate regression spline documentation to GAM algorithm page.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7656'>#7656</a>] - Added missing parameters to XGBoost algorithm page.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7652'>#7652</a>] - Added more information about log files to User Guide.
</li>
</ul>
Zermelo (3.32.0.5) - 3/16/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/5/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7844'>#7844</a>] - GAM no longer creates multiple knots at the same coordinates when the cardinality of the `gam_columns` is less than the number of `knots` specified by the user.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7694'>#7694</a>] -
Feature interactions can now be save as .xlxs files.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7614'>#7614</a>] - Job polling will retry connecting to h2o nodes if connection fails.
</li>
</ul>
Zermelo (3.32.0.4) - 2/1/2021
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/4/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7697'>#7697</a>] - Partial Dependence Plot no longer failing for High Cardinality even when `user_splits` is defined.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7695'>#7695</a>] - Fixed failing Delta Lake import for Python API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7686'>#7686</a>] - Fix Stacked Ensemble’s incorrect handling of fold column.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7902'>#7902</a>] - Added MOJO support for CoxPH.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7675'>#7675</a>] - Escape all quotes by default when writing CSV.
</li>
</ul>
Docs
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7701'>#7701</a>] - Added to docs that AUCPR can be plotted.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7684'>#7684</a>] - Updated the Customer Algorithm graphic for the Architecture section of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7661'>#7661</a>] - Updated the copyright year to 2021.
</li>
</ul>
Zermelo (3.32.0.3) - 12/24/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/3/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7868'>#7868</a>] - The `pca_impl` parameter is no longer passed to PCA MOJO.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7749'>#7749</a>] - Objects to be retained no longer removed during the `h2o.removeAll()` command.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7743'>#7743</a>] - Starting GridSearch in a fresh cluster with new hyperparameters that overlap old ones will no longer cause the old models to be trained again.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7731'>#7731</a>] - GridSearch no longer hangs indefinitely when not using the default value for paralellism.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7724'>#7724</a>] - Fixed the parent dir lookup for HDFS grid imports.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7717'>#7717</a>] - Fixed the CustomDistribution test error.
</li>
</ul>
New Feature
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12775'>#12775</a>] - Cross-Validation predictions can now be saved alongside the model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8366'>#8366</a>] - Added multinomial and grid search support for AUC/PR AUC metrics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7782'>#7782</a>] - Now offers a standalone R client that doesn’t include the h2o jar.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7773'>#7773</a>] - Created a Red Hat certification for H2O Docker Image.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7764'>#7764</a>] - Fixed randomized split points for `histogram_type=“Random”` when nbins=2.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7729'>#7729</a>] - Single quote regime for CSV parser exposed for importing & uploading files.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7887'>#7887</a>] - REST API disabled on non-leader Kubernetes nodes.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7769'>#7769</a>] - GLM now uses proper logging instead of printlines.
</li>
</ul>
Docs
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7819'>#7819</a>] -
Added non-tree-based models to the variable importance page in the user guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7775'>#7775</a>] - Updated the AutoML citation in the User Guide to point to the H2O AutoML ICML AutoML workshop paper.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7762'>#7762</a>] - Updated Python docstring examples about cross-validation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7740'>#7740</a>] - Corrected `k` parameter description for PCA.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7723'>#7723</a>] - Corrected the RuleFit Python example.
</li>
</ul>
Zermelo (3.32.0.2) - 11/17/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/2/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7849'>#7849</a>] - Implemented deserialization of monotone constraints.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7798'>#7798</a>] - Updated required version of ggplot2 in R package to 3.3.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7778'>#7778</a>] - Fixed the parsing of GLM’s `rand_family` params in MOJO JSON.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7768'>#7768</a>] - Fixed NPE that resulted when starting a grid with SequentialWalker in AutoML exploitation phase.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7765'>#7765</a>] - Fixed MOJO version check message.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7759'>#7759</a>] - When grid search has parallelism enabled, it now includes CV models.
</li>
</ul>
New Feature
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7900'>#7900</a>] - Added feature interactions and importance for XGBoost and GBM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7867'>#7867</a>] - Added new `interaction_constraints` parameter to XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7804'>#7804</a>] - Added an option to not have quotes in the header during exportFile.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7758'>#7758</a>] - Added ability to retrieve a list of all the models in an H2O cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7745'>#7745</a>] - Added custom pod labels for HELM charts.
</li>
</ul>
Task
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7807'>#7807</a>] - Added `lambda_min` & `lambda_max` parameters to GLMModelOutputs.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7894'>#7894</a>] - Added default values to all algorithm parameters in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7890'>#7890</a>] - Fixed the discrepancies between the Target Encoding User Guide page and Client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7808'>#7808</a>] - Added ONNX support to the documentation.
</li>
</ul>
Engineering Story
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7796'>#7796</a>] - Added a new method which properly locks H2O Frames during conversion from Spark Data Frames to H2O Frames in Sparkling Water.
</li>
</ul>
Docs
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7806'>#7806</a>] - On the Grid Search User Guide page, fixed the missing syntax highlight in the Python example of the Random Grid Search section.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7805'>#7805</a>] - Added `rule_generation_ntrees` parameter to the RuleFit page.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7767'>#7767</a>] - Added documentation for GBM and XGBoost on feature interactions and importance.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7757'>#7757</a>] - Added a Python example to the `stratify_by` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7747'>#7747</a>] - Added a Feature Engineering section to the Data Manipulation page in the User Guide.
</li>
</ul>
Zermelo (3.32.0.1) - 10/8/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zermelo/1/index.html</a>
Bug
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7972'>#7972</a>] - Fixed StackedEnsemble’s retrieval of the seed parameter value.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7893'>#7893</a>] - Deserialization values of MOJO ModelParameter now work when the Value Type is int[].
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7881'>#7881</a>] - H2O no longer uses lazy-loading for sequential zip parse.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7879'>#7879</a>] - Updated model_type argument names for Rulefit in R.
</li>
</ul>
New Feature
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8393'>#8393</a>] - Quantile distributions added to monotone constraints.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8318'>#8318</a>] - TargetEncoder integrated into ModelBuilder.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7885'>#7885</a>] - Python client no longer instructs the user to declare a root handler in library mode.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7851'>#7851</a>] - Hostname used as certificate alias to lookup machine-specific certificate allowing Hadoop users to connect to Flow over HTTPS.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7846'>#7846</a>] - Added the model explainability interface for H2O models and AutoML objects in both R & Python.
</li>
<li>
[<a href='https://github.com/h2oai/h2o-3/issues/7919'>#7919</a>] - Added the RuleFit algorithm for interpretability.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7834'>#7834</a>] - Implemented a basic HELM chart.
</li>
</ul>
Task
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7878'>#7878</a>] - Rulefit model added to algorithm section of UserGuide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7855'>#7855</a>] - Added an Explainability page to the User Guide outlining the new `h2o.explain()` and `h2o.explain_row()` functions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7838'>#7838</a>] - Updated the AutoML User Guide page to include the new Explainability and Preprocessing sections.
</li>
</ul>
Improvement
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12783'>#12783</a>] - Added support for Python 3.7+.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7922'>#7922</a>] - Exposes names of score0 output values in MOJO.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7909'>#7909</a>] - Added function to plot a Precision Recall Curve.
</li>
<li>
[<a href='https://github.com/h2oai/h2o-3/issues/7899'>#7899</a>] - RuleFit model represented by the set of rules obtained from trees during training.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7876'>#7876</a>] - Performance improved for exporting a Frame to CSV.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7872'>#7872</a>] - GPU backend allowed in XGBoost when running multinode even when `build_tree_one_node` is enabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7863'>#7863</a>] - Updated all URLs in R package to use HTTPS.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7852'>#7852</a>] - Upgraded to XGBoost 1.2.0.
</li>
</ul>
Technical task
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8271'>#8271</a>] - Added cross-validation to GAM allowing users to find the best alpha/lambda values when building a GAM model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7967'>#7967</a>] - Added TargetEncoder support for multiclass problems.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7896'>#7896</a>] - Added new TargetEncoder parameter that allows users to remove original features automatically.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7862'>#7862</a>] - Implemented minimal support for TargetEncoding in AutoML.
</li>
</ul>
Docs
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8097'>#8097</a>] - Updated the descriptions of AutoML in R & Python packages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7860'>#7860</a>] - Made the default for `categorical_encoding` in XGBoost explicit in the documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7831'>#7831</a>] - Updated the import datatype section of the Python FAQ in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7826'>#7826</a>] - Updated the default values for `min_rule_length` and `max_rule_length` on the RuleFit page of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7825'>#7825</a>] - Updated the `validation_frame` definition for unsupervised algorithms in the User Guide.
</li>
</ul>
Zeno (3.30.1.3) - 9/28/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zeno/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zeno/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7861'>#7861</a>] - CRAN - Use HTTPS for all downloads within the R package.
</li>
</ul>
Zeno (3.30.1.2) - 9/3/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zeno/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zeno/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11601'>#11601</a>] - The ‘h2o.unique()’ command will now only return the unique values within a column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8012'>#8012</a>] - k-LIME easy predict wrapper now uses Regression or KLime as a model category instead of just KLime.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7982'>#7982</a>] - Fixed the CRAN check warnings on r-devel for cross-references in the R documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7935'>#7935</a>] - Documentation added detailing the supported encodings for CSV files.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7931'>#7931</a>] - GLM parameters integrated into GAM parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7901'>#7901</a>] - Fixed broken URLs in R documentation that caused CRAN failures.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8017'>#8017</a>] - Added the concordance statistic for CoxPH models.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8157'>#8157</a>] - When using multiple alpha/lambda values for calling GLM from GAM, GLM now returns the best results across all alpha/lambda values. Also added the ‘cold_start’ parameter added to GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7918'>#7918</a>] - Added documentation for new GAM hyperparameter ’subspaces’.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7913'>#7913</a>] - GLM new parameter ‘cold_start’ added to User Guide and GLM booklet.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7985'>#7985</a>] - Reduced the memory cost of the `drop_duplicate` operation by cleaning up data early.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7916'>#7916</a>] - When calculating unique() values on a column that is the result of an AstRowSlice operation, the domain is now collected in-place and no longer results in an error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7908'>#7908</a>] - Categorical encoding documentation updated by adding ‘EnumLimited’ & ’SortByReponse’ to KMeans and removing ‘Eigen’ from XGBoost.
</li>
</ul>
<h4>Technical task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8270'>#8270</a>] - Tests added to verify grid search functionality for GAM and allows the user to create more complex hyper spaces for grid search by adding ‘subspaces’ key and functionality to grid search backend.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7906'>#7906</a>] - Added documentation on how to retrieve reproducibility information.
</li>
</ul>
Zeno (3.30.1.1) - 8/10/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zeno/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zeno/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8521'>#8521</a>] - H2OFrames with fields containing double quotes/line breaks can now be converted to Pandas dataframe.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8149'>#8149</a>] - Impossible to set Max_depth to unlimited on DRF classifer
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8004'>#8004</a>] - Model generation for MOJO/POJO are disabled when interaction columns are used in GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7993'>#7993</a>] - Reproducibility Information Table now hidden in H2O-Flow.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11793'>#11793</a>] - Added support for `offset_column` in the Stacked Ensemble metalearner.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11794'>#11794</a>] - Added support for `weights_column` in the Stacked Ensemble metalearner.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8826'>#8826</a>] - Added continued support to Generalized Additive Models for H2O.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8397'>#8397</a>] - The value of model parameters can be retrieved at the end of training, allowing users to retrieve an automatically chosen value when a parameter is set to AUTO.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8353'>#8353</a>] - H2O Frame is now able to be saved into a Hive table.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8171'>#8171</a>] - XGBoost can now be executed on an external Hadoop cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7999'>#7999</a>] - Added the `contamination` parameter to Isolation Forest which is used to mark anomalous observations.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7998'>#7998</a>] - Introduced the `validation_response_column` parameter for Isolation Forest which allows users to name the response column in the validation frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7992'>#7992</a>] - Added official support for Java 14 in H2O.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7942'>#7942</a>] - Added external cluster startup timeout for XGBoost.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7990'>#7990</a>] - Hadoop Docker image run independent of S3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7966'>#7966</a>] - Upgraded the build/test environment to support R 4.0 and Roxygen2.7.1.1.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8698'>#8698</a>] - Implemented TF-IDF algorithm to reflect how important a word is to a document or collection of documents.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8691'>#8691</a>] - GridSearch R API test added for Isolation Forest.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8193'>#8193</a>] - ‘AUTO’ option added for GLM & GAM family parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8142'>#8142</a>] - XGBoost Variable Importances now computed using a Java predictor.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8091'>#8091</a>] - StackedEnsemble can now be created using only monotone models if user specifies `monotone_constraints` in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8071'>#8071</a>] - Enabled using imported models as base models in Stacked Ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7988'>#7988</a>] - Removed deprecated H2O-Scala module.
</li>
</ul>
<h4>Technical Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8447'>#8447</a>] - Added Java backend to support MOJO in GAM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8028'>#8028</a>] - Added support for `early_stopping` parameter in GAM and GLM.
</li>
</ul>
<h4> Engineering Story
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7938'>#7938</a>] - Sparkling Water Booklet removed from the H2O-3 repository.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8082'>#8082</a>] - Added H2O Client chapter to the User Guide which includes section on Sklearn integration.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8000'>#8000</a>] - Added documentation in the Isolation Forest section for the `contamination` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7991'>#7991</a>] - Added documentation in GLM & GAM, and the `family` & `link` algorithm parameters to include how `family` can now be set equal to AUTO.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7984'>#7984</a>] - Added `gains lift_bins` to the parameter appendix and added and example to the parameter in the Python documentation. Added an example for the Kolmogorov-Smirnov metric to the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7983'>#7983</a>] - Updated GAM and GLM documentation to include support for `early_stopping`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7978'>#7978</a>] - Added the Kolmogorov-Smirnov metric formula to the Performance and Prediction chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7960'>#7960</a>] - Added the `negativebinomial` value to the `family` parameter page.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7959'>#7959</a>] - Added the `ordinal` and `modified_huber` values to the `distribution` parameter page.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7957'>#7957</a>] - Updated deprecated parameter `loading_name` to `representation_name` and fixed the broken init link in the GLRM section of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7955'>#7955</a>] - Added a note in the User Guide Stacked Ensemble section about building a monotonic Stacked Ensemble.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7940'>#7940</a>] - Added documentation for how `balance_classes` is triggered.
</li>
</ul>
Zahradnik (3.30.0.7) - 7/21/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/7/index.html</a>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8207'>#8207</a>] - Added support for partitionBy column in partitioned parquet or CSV files.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7994'>#7994</a>] - Warning added for user if both a lamba value and lambda search are provided in GLM.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12662'>#12662</a>] - Added `max_runtime_secs` parameter to Stacked Ensemble.
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/28'>private-#28</a>] - Upgraded Jetty 9 and switched default webserver to Jetty 9.
</li>
</ul>
Zahradnik (3.30.0.6) - 6/30/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/6/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8009'>#8009</a>] - GLM Plug values are now propagated to MOJOs/POJOs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8008'>#8008</a>] - In the Python documentation, the HGLM example now references `random_columns` by indices rather than by column name.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/7997'>#7997</a>] - Fixed a link to H2O blogs in the R documentation.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8233'>#8233</a>] - Added support for the Kolmogorov-Smirnov metric for binary classification models.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8014'>#8014</a>] - Added documentation in the Performance and Prediction chapter for the Kolmogorov-Smirnov metric.
</li>
</ul>
Zahradnik (3.30.0.5) - 6/18/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8308'>#8308</a>] - Fixed an issue that denied all requests to display H2O Flow in an iframe.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8075'>#8075</a>] - Importing with `use_temp_table=False` now works correctly on Teradata.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8050'>#8050</a>] - Building a GLM model with `interactions` and `lambda = 0` no longer produces a "Categorical value out of bounds" error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8048'>#8048</a>] - Fixed an inconsistency that occurred when using `predict_leaf_node_assignment` with a path and with a terminal node. For trees with a max_depth of up to 63, the results now match. For max_depth of 64 or higher (for path and nodes that are "too deep"), H2O will no longer produce incorrect results. Instead it will return "NA" for tree paths and "-1" for node IDs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8042'>#8042</a>] - Leaf node assignment now works correctly for trees with a depth >= 31. Note that for trees with a max_depth of 64 or higher, H2O will return "NA" for tree paths and "-1" for node IDs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8039'>#8039</a>] - `allow_insecure_xgboost` now works correctly on Hadoop.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8206'>#8206</a>] - HTML documentation is now available as a downloadable zip file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8037'>#8037</a>] - Users can now retrieve the prediction contributions when running `mojo_predict_pandas` in Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8025'>#8025</a>] - H2O documentation is now available in an h2odriver distribution zip file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8024'>#8024</a>] - Quantiles models during the training of other models are now recognized as a regular model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8018'>#8018</a>] - The H2O-SCALA module is deprecated and will be removed in a future release.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9202'>#9202</a>] - Added support for models built with any `family` when running makeGLMModel.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8052'>#8052</a>] - K8S Docker images for h2o-3 are now available.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8023'>#8023</a>] - Warnings are now produced during model building when using the Python client.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8495'>#8495</a>] - Added examples for saving and loading grids in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8051'>#8051</a>] - Improved the examples in the Performance and Prediction chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8049'>#8049</a>] - In the AutoML Random Grid Search Parameters topic, removed the no-longer-supported `min_sum_hessian_in_leaf` parameter from the XGBoost table. Also added clarification on how GHL models are handled in an AutoML random grid search run.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8035'>#8035</a>] - In the Python documentation, add examples for Grid Metrics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8013'>#8013</a>] - The value of T as described in the description for `categorical_encoding="enum_limited"` is 10, not 1024.
</li>
</ul>
Zahradnik (3.30.0.4) - 6/1/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8275'>#8275</a>] - h2o.merge() now works correctly when you joining an H2O frame where the join is on a <dbl> column to another frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8184'>#8184</a>] - Fixed an issue that caused h2o.get_leaderboard to fail after creating an AutoML object, disconnecting the client, starting a new session, and then reconecting to the running H2O cluster for the re-attached H2OAutoML object.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8147'>#8147</a>] - Stacked Ensemble now inherits distributions/families supported by the metalearner.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8137'>#8137</a>] - Fixed an issue that caused AutoML to fail when the target included special characters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8073'>#8073</a>] - CAcert is now supported with the Python API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8069'>#8069</a>] - Water Meter and Form Login now work correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8066'>#8066</a>] - In Aggregator, added support for retrieving the Mappings Frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8056'>#8056</a>] - Added support for using monotone constraints with Tweedie distribution in GBM.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10207'>#10207</a>] - Added a new drop_duplicates function to drop duplicate observations from an H2O frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9371'>#9371</a>] - Partial dependence plots are now available for multiclass problems.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8134'>#8134</a>] - Users now receive a warning if they try to get variable importances in Stacked Ensemble.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8111'>#8111</a>] - In XGBoost, removed the min_sum_hessian_in_leaf and min_data_in_leaf options, which are no longer supported by XGBoost. Also added the `colsample_bynode` option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8089'>#8089</a>] - data.table warning messages are now suppressed inside h2o.automl() in R.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8120'>#8120</a>] - Added a "Training Models" section to the User Guide, which describes train() and train_segments().
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8113'>#8113</a>] - Updated XGBoost to indicate that this version requires CUDA 9, and included information showing users how to check their CUDA version.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8112'>#8112</a>] - Added information about GAM support to the missing_values_handling parameter appendix entry.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8107'>#8107</a>] - Updated the Minio Instance topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8064'>#8064</a>] - `monotone_constraints` can now be used with `distribution=tweedie`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8062'>#8062</a>] - Updated the PDP topic to include support for multinomial problems and updated the examples.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8053'>#8053</a>] - In the API-related Changes topic, noted that `min_sum_hessian_in_leaf` and `min_data_in_leaf` are no longer supported in XGBoost.
</li>
</ul>
Zahradnik (3.30.0.3) - 5/12/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8146'>#8146</a>] - Improved validation and error messages for CoxPH.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8140'>#8140</a>] - In XGBoost, the `predict_leaf_node_assignment` parameter now works correctly with multiclass.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8121'>#8121</a>] - Fixed an issue that caused GBM to fail when it encountered a bin that included a single value and the rest NAs.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8537'>#8537</a>] - Updated the AutoML example in the R package.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8199'>#8199</a>] - PDPs now allow y-axis scaling options.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8175'>#8175</a>] - Improved speed for training and prediction of Stacked Ensembles.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12851'>#12851</a>] - Added tables showing parameter values and random grid space ranges to the AutoML chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8294'>#8294</a>] - Improved the Hive import documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8138'>#8138</a>] - Improved documentation for Quantiles in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8133'>#8133</a>] - Fixed the documented default value for `min_split_improvement` parameter in XGBoost.
</li>
</ul>
Zahradnik (3.30.0.2) - 4/28/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8237'>#8237</a>] - Fixed an issue that caused H2O to crash while debugging Python code using intellij/pycharm.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8211'>#8211</a>] - Fixed an issue that caused an assertion error while running Grid Search.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8203'>#8203</a>] - Training of a model based on a data frame that includes Target Encodings no longer fails due to a locked frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8198'>#8198</a>] - Added train_segments() to the R html documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8196'>#8196</a>] - Target Encoder now unlocks the output frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8182'>#8182</a>] - Fixed the BiasTerm in XGBoost Contributions after upgrading to XGBoost 1.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8152'>#8152</a>] - GBM and XGBoost no longer ignore a column that includes a constant and NAs.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8284'>#8284</a>] - Added the following options for customizing and retrieving threshold values.
<ul>
<li>`threshold` allows you to specify the threshold value used for calculating the confusion matrix.
<li>`default_threshold` allows you to change the threshold that is used to binarise the predicted class probabilities.</li>
<li>`reset_model_threshold` allows you to reset the model threshold.</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8261'>#8261</a>] - Introduced Kubernetes integration. Docker image tests are now available on K8S and published to Docker Hug.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8229'>#8229</a>] - A progress bar is now available during Shap Contributions calculations.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9209'>#9209</a>] - An H2O Frame containing weights can now be specified when running `make_metrics`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8361'>#8361</a>] - Added POJO and MOJO support for all encodings in GBM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8192'>#8192</a>] - Users will now receive an error if they attempt to run https in h2o.init() when starting a local cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8173'>#8173</a>] - Added an `-allow_insecure_xgboost` option to h2o and h2odriver that allows XGBoost multinode to run in a secured cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8169'>#8169</a>] - Only the leader node is exposed on K8S.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8620'>#8620</a>] - Updated the Target Encoding topic and examples based on the improved API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8293'>#8293</a>] - Added a new "Supported Data Types" topic to the Algorithms chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8195'>#8195</a>] - Added a new "Kubernetes Integration" topic to the Welcome chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8194'>#8194</a>] - Fixed the links for the constrained k-means Python demos.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8191'>#8191</a>] - Fixed the R example in the GAM chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8187'>#8187</a>] - Added clarification for when `min_mem_size` and `max_mem_size`` are set to NULL/None in h2o.init().
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8179'>#8179</a>] - The link to the slideshare in the DRF chapter now points to https instead of http.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8177'>#8177</a>] - Added information about the h2o.get_leaderboard() function to the AutoML chapter of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8176'>#8176</a>] - Updated the MOJO Quickstart showing how to use PrintMojo to visualize MOJOs without requiring Graphviz.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8168'>#8168</a>] - The import_mojo() function now uses "path" instead of "dir" when downloading, importing, uploading, and saving models. Updated the examples in the documentation.
</li>
</ul>
Zahradnik (3.30.0.1) - 4/3/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-zahradnik/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8638'>#8638</a>] - Fixed an issue that caused performing multiple h2o.init() to fails with R on Windows.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8545'>#8545</a>] - Increased the default clouding time to avoid times out that resulted in a Cloud 1 under 4 error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8296'>#8296</a>] - Removed obsolete exactLambdas parameter from GLM.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12884'>#12884</a>] - Added support for a fractional response in GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8826'>#8826</a>] - Added support for Generalized Additive Models (GAMs) in H2O. The documentation for this newly added algorithm can be found <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/gam.html'>here</a>.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8730'>#8730</a>] - Added support for parallel training (e.g. spark_apply in rsparkling or Python/R).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8404'>#8404</a>] - Added support for Continuous Bag of Words (CBOW) models in Word2Vec.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8369'>#8369</a>] - H2O can now predict OOME during parsing and stop the job if OOME is imminent.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8333'>#8333</a>] - Add GBM POJO support for SortByResponse and enumlimited.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8290'>#8290</a>] - Added support for Leaf Node Assignments in XGBoost and Isolation Forest MOJOs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8285'>#8285</a>] - Added support for importing Stacked Ensemble MOJO models for scoring. (Note that this only applies to Stacked Ensembles that include algos with MOJO support.)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8232'>#8232</a>] - Added support for the `single_node_mode` parameter in CoxPH.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8228'>#8228</a>] - H2O now provides the original algorithm name for MOJO import.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8215'>#8215</a>] - Created a segmented model training interface in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8214'>#8214</a>] - Added a print method for the H2OSegmentModel object type in R.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8401'>#8401</a>] - Removed the previously deprecated DeepWater Estimator function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8252'>#8252</a>] - Now using Java-based scoring for XGBoostModels.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11521'>#11521</a>] - In the H2O R package, `data.table` is now enabled by default (if installed).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9327'>#9327</a>] - In AutoML, users can try tuning the learning rate for the best model found during exploration in XGBoost and GBM. Note that the new `exploitation_ratio` parameter is still experimental.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8781'>#8781</a>] - Added out-of-the-box support for starting an h2o cluster on Kubernetes. Refer to this <a href='https://github.com/h2oai/h2o-3/blob/master/h2o-k8s/README.md'>README</a> for more information.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8553'>#8553</a>] - Improved the way AUC-PR is calculated.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8430'>#8430</a>] - Added an option to upload binary models from Python and R.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8564'>#8564</a>] - Added examples for Grid Search in the Python Module documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8481'>#8481</a>] - Added examples to the R Reference Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8287'>#8287</a>] - Added documentation for the fractional binomial family in the GLM section.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8286'>#8286</a>] - Added documentation for the new GAM algorithm.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8249'>#8249</a>] - Updated tab formatting for the `cluster_size_constraints` parameter appendix entry.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8231'>#8231</a>] - Updated the Target Encoding R example.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8230'>#8230</a>] - Included confusion matrix threshold details for binary and multiclass classification in the Performance and Prediction chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8227'>#8227</a>] - Added documentation for new `upload_model` function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8221'>#8221</a>] - Improved documentation around citing H2O in publications.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8209'>#8209</a>] - Added documentation for `single_node_mode` in CoxPH.
</li>
</ul>
Yule (3.28.1.3) - 4/2/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yule/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yule/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8300'>#8300</a>] - Fixed an issue that occurred during Hive SQL import with `fetch_mode=SINGLE`; improved Hive SQL import speed; added an option to specify the number of chunks to parse.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8251'>#8251</a>] - Hive delegation token refresh now recognizes `-runAsUser`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8243'>#8243</a>] - Fixed `base_model` selection for Stacked Ensembles in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8241'>#8241</a>] - The Parquet parser now supports arbitrary precision decimal types.
</li>
</ul>
<h4>Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8246'>#8246</a>] - The H2O Hive parser now recognizes varchar column types.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8223'>#8223</a>] - Hive tokens are now refreshed without distributing the Steam keytab.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8469'>#8469</a>] - Users can now specify the `max_log_file_size` when starting H2O. The log file size currently defaults to 3MB.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8279'>#8279</a>] - Fixed the of parameters for TargetEncoder in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8247'>#8247</a>] - HostnameGuesser.isInetAddressOnNetwork is now public.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8235'>#8235</a>] - Improved mapper-side Hive delegation token acquisition. Now when H2O is started from Steam, the Hive delegation token will already be acquired when the cluster is up.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8257'>#8257</a>] - Added to docs that `transform` only works on numerical columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8218'>#8218</a>] - Added documentation for the new num_chunks_hint option that can be specified with `import_sql_table`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8217'>#8217</a>] - Added documentation for the new `max_log_file_size` H2O starting parameter.
</li>
</ul>
Yule (3.28.1.2) - 3/17/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yule/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yule/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8847'>#8847</a>] - The `base_models` attribute in Stacked Ensembles is now populated in both Python and R.
<br/>
Note that in Python, if there are no `base_models` in `_parms`, then `actual_params` is used to retrieve base_models, and it contains the names of the models. In R, `ensemblemodel$base_models` is populated with a vector of base model names.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8344'>#8344</a>] - Fixed an issue that caused the leader node to be overloaded when parsing 30k+ Parquet files.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8332'>#8332</a>] - Fixed an issue that caused `model end_time` and `run_time` properties to return a value of 0 in client mode.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8280'>#8280</a>] - TargetEncoderModel's summary no longer prints the fold column as a column that is going to be encoded by this model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8273'>#8273</a>] - When h2omapper fails before discovering SELF (ip & port), the log messages are no longer lost.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8506'>#8506</a>] - Added DeepLearning MOJO support in Generic Models.
</li>
</ul>
<h4>Improvement\</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9031'>#9031</a>] - Changed the output format of `get_automl` in Python from a dictionary to an object.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8266'>#8266</a>] - Users can now specify `-hdfs_config` multiple times to specify multiple Hadoop config files.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8264'>#8264</a>] - Fixed an issue that caused the clouding process to time out for the Target Encoding module and resulted in a `Cloud 1 under 4` error.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8445'>#8445</a>] - Improved FAQ describing how to use the H2O-3 REST API from Java.
</li>
</ul>
Yule (3.28.1.1) - 3/5/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yule/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yule/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8314'>#8314</a>] - Added missing AutoML global functions to the Python and R documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8312'>#8312</a>] - In the Python client, improved the H2OFrame documentation and properly labeled deprecated functions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8303'>#8303</a>] - Fixed an issue that caused imported MOJOs to produce different predictions than the original model.
</li>
</ul>
<h4>Engineering Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8310'>#8310</a>] - Removed Sparling Water external backend code from H2O.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8309'>#8309</a>] - In the R client docs for h2o.head() and h2o.tail(), added an example showing how to control the number of columns to display in dataframe when using a Jupyter notebook with the R kernel.
</li>
</ul>
Yu (3.28.0.4) - 2/23/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yu/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yu/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9016'>#9016</a>] - DeepLearning MOJOs are now thread-safe.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8406'>#8406</a>] - Fixed an issue that caused h2oframe.apply to fail when run in Python 3.7. Note that Python 3.7 is still not officially supported, but support is a WIP.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8375'>#8375</a>] - XGBoost now correctly respects monotonicity constraints for all tree_methods.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8373'>#8373</a>] - Decision Tree descriptions no longer include more descriptions than `max_depth` splits.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8365'>#8365</a>] - Fixed an issue that caused `import_hive_table` to fail with a JDBC source and a partitioned table.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8364'>#8364</a>] - Improved the DKVManager sequential removal mechanism.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8356'>#8356</a>] - In XGBoost, added a message indicating that the `exact` tree method is not supported in multinode.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8329'>#8329</a>] - XGBoost ContributionsPredictor is now serializable.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8328'>#8328</a>] - Fixed a CRAN warning related to ellipsis within arguments in the R package.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8325'>#8325</a>] - Added support for specifying AWS session tokens.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9179'>#9179</a>] - Added support for Constrained K-Means clustering.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8674'>#8674</a>] - In Stacked Ensembles, added support for "xgboost" and "naivebayes" in the `metalearner_algorithm` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8334'>#8334</a>] - Added support for `build_tree_one_node` in XGBoost.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8507'>#8507</a>] - In the R client, users can now optionally specify the number of columns to display in `h2o.frame`, `h2o.head`, and `h2o.tail`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8443'>#8443</a>] - Fixed an issue that caused AutoML to fail to run if XGBoost was disabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8382'>#8382</a>] - Stacktraces are no longer returned in `h2o.getGrid` when failed models are present.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8327'>#8327</a>] - Added `createNewChunks` with a "sparse" parameter in ChunkUtils.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8675'>#8675</a>] - Added an FAQ to the MOJO and POJO quick starts noting that MOJOs and POJOs are thread safe for all supported algorithms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8420'>#8420</a>] - Added the new `cluster_size_constraints` parameter to the KMeans chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8350'>#8350</a>] - Updated docs to specify that `mtries=-2` gives all features.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8323'>#8323</a>] - Updated EC2 and S3 Storage topic to include the new, optional AWS session token.
</li>
</ul>
Yu (3.28.0.3) - 2/5/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yu/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yu/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8888'>#8888</a>] - In the R client, fixed a parsing bug that occurred when using quotes with .csv files in as.data.frame().
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8815'>#8815</a>] - Fixed an Unsupported Operation Exception in UDP-TCP-SEND.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8522'>#8522</a>] - GLM now supports coefficients on variable importance when model standardization is disabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8446'>#8446</a>] - In the Python client, rbind() can now be used on all numerical types.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8440'>#8440</a>] - In XGBoost, fixed an error that occurred during model prediction when OneHotExplicit was specified during model training.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8429'>#8429</a>] - Performing grid search over Target Encoding parameters now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8391'>#8391</a>] - Fixed an issue that caused import_hive_table to not classload the JDBC driver.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8389'>#8389</a>] - MOJOs can now be built from XGBoost models built with an offset column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8363'>#8363</a>] - Fixed an issue that cause the R and Python clients to return the wrong sensitivity metric value.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8362'>#8362</a>] - Fixed an incorrect sender port calculation in TimestampSnapshot.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9128'>#9128</a>] - In AutoML, multinode XGBoost is now enabled by default.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8410'>#8410</a>] - Users can now specify a custom JDBC URL to retrieve the Hive Delegation token using hiveJdbcUrlPattern.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8385'>#8385</a>] - In XGBoost fixed a deprecation warning for reg:linear.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8442'>#8442</a>] - import_folder() can now be used when running H2O in GCS.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8407'>#8407</a>] - Added support for registering custom servlets.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8377'>#8377</a>] - In XGBoost, when a parameter with a synonym is updated, the synonymous parameter is now also updated.
</li>
</ul>
<h4>Engineering Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8388'>#8388</a>] - AutoBuffer.getInt() is now public.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8412'>#8412</a>] - Python examples for plot method on binomial models now use the correct method signature.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8411'>#8411</a>] - Updated custom_metric_func description to indicate that it is not supported in GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8395'>#8395</a>] - Updated the AutoML documentation to indicate that multinode XGBoost is now turned on by default.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8379'>#8379</a>] - Fixed the description for the Hadoop -nthreads parameter.
</li>
</ul>
Yu (3.28.0.2) - 1/20/2020
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yu/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yu/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8571'>#8571</a>] - Fixed an issue that resulted in a "DistributedException java.lang.ClassNotFoundException: BAD" message.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8499'>#8499</a>] - Users can now specify either a model or a model key when checkpointing.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8490'>#8490</a>] - Fixed an issue that resulted in an endless loop when CsvParser parser $ sign was enclosed in quotes.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8480'>#8480</a>] - In GBM and DRF, fixed an AIOOBE error that occurred when the dataset included negative zeros (-0.0).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8467'>#8467</a>] - Fixed a race condition in the addWarningP method on Model class.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8455'>#8455</a>] - h2odriver now gets correct version of Hadoop dependencies.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8439'>#8439</a>] - Fixed a race condition in addVec.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8435'>#8435</a>] - Parallel Grid Search threads now call the Hyperspace iterator one at a time.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8431'>#8431</a>] - sklearn wrappers now expose wrapped estimator as a public property.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8428'>#8428</a>] - Fixed an issue in reading user_splits in Java.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8421'>#8421</a>] - Fixed an issue that caused rank vectors of Spearman correlation to have different chunk layouts.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8583'>#8583</a>] - Added a JSON option of PrintMojo.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8520'>#8520</a>] - Improved the error message that displays when a user attempts to import data from an HDFS directory that is empty.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8456'>#8456</a>] - H2O can now read Hive table metadata two ways: either via direct Metastore access or via JDBC.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9167'>#9167</a>] - Improved heuristics used for finding IP addresses on Hadoop in order to select the right subnet automatically.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8605'>#8605</a>] - Added support for `offset_column in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8551'>#8551</a>] - Users can now create tree visualizations without installing additional packages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8503'>#8503</a>] - Added a new `download_model` function for downloading binary models in the R and Python clients.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8475'>#8475</a>] - Improved XGBoost performance.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8474'>#8474</a>] - When computing the correlation matrix of one or two H2OFrames (using `cor()`), users can now specify a method of either Pearson (default) or Spearman.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8438'>#8438</a>] - Users are now warned when they attempt to run AutoML with a validation frame and with nfolds > 0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8436'>#8436</a>] - AutoML no longer trains a "Best of Family Stacked Ensemble" when only one family is specified.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12973'>#12973</a>] - Removed `ignored_columns` from the list of available paramters in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8647'>#8647</a>] - Fixed a broken link in the JAVA FAQ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8552'>#8552</a>] - Improved the documentation for Tree Class in the Python Client docs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8484'>#8484</a>] - Clarified the difference between h2o.performance() and h2o.predict() in the Performance and Prediction chapter of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8478'>#8478</a>] - Incorporated HGLM documentation updates into the GLM booklet.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8441'>#8441</a>] - Added an FAQ for GC allocation failure in the FAQ > Clusters section.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8434'>#8434</a>] - In the Stacked Ensembles chapter, improved the metalearner support FAQ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8419'>#8419</a>] - Added `offset_column` to the list of supported parameters in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8418'>#8418</a>] - Added information about recent API changes in AutoML to the <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/api-changes.html">API-Related Changes</a> section in the User Guide.
</li>
</ul>
Yu (3.28.0.1) - 12/16/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yu/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yu/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12823'>#12823</a>] - AutoML reruns using, for example, the same project name, no project name, etc., now produce consistent results.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8924'>#8924</a>] - Fixed an issue that occcurred when running an AutoML instance twice using the same project_name. AutoML no longer appends new models to the existing leaderboard, which caused the models for the first run to attempt to get rescored against the new learderboard_frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8696'>#8696</a>] - Updated the list of stopping metric options for AutoML in Flow. Also added support for the aucpr stopping metric in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8673'>#8673</a>] - When training a K-Means model, the framename is no longer missing in the training metrics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8642'>#8642</a>] - In AutoML, the `project_name` is now restricted to the same constraints as h2o frames.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8576'>#8576</a>] - In GBM, fixed an NPE that occurred when sample rate < 1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8575'>#8575</a>] - The AutoML backend no longer accepts `ignored_columns` that contain one of response column, fold column, or weights column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8505'>#8505</a>] - XGBoost MOJO now works correctly in Spark.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8504'>#8504</a>] - The REST API ping thread now starts after the cluster is up.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8502'>#8502</a>] - Fixed an NPE at hex.tree.TreeHandler.fillNodeCategoricalSplitDescription(TreeHandler.java:272)
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12219'>#12219</a>] - Extended MOJO support for PCA
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9121'>#9121</a>] - We are very excited to add HGLM (Hierarchical GLM) to our open source offering. As this is the first release, we only implemented the Gaussian family. However, stay tuned or better yet, tell us what distributions you want to see next. Try it out and send us your feedback!
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9117'>#9117</a>] - MOJO Import is now available for XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8917'>#8917</a>] - Improved integration of the H2O Python client with Sklearn.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8896'>#8896</a>] - Users can now specify monotonicity constraints in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8884'>#8884</a>] - Users can now save and load grids to continue a Grid Search after a cluster restart.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8860'>#8860</a>] - Users can now specify a `parallelism` parameter when running grid search. A value of 1 indicagtes sequential building (default); a value of 0 is used for adapative parallelism; and any value greater than 1 sets the exact number of models built in parallel.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8837'>#8837</a>] - Added a function to calculate Spearman Correlation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8793'>#8793</a>] - Users can now specify the order in which training steps will be executed during an AutoML run. This is done using the new `modeling_plan` option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8745'>#8745</a>] - The `calibration_frame` and `calibrate_model` options can now be spcified in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8707'>#8707</a>] - Added support for OneHotExplicit categorical encoding in EasyPredictModelWrapper.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8568'>#8568</a>] - Added aucpr to the AutoML leaderboard, stopping_metric, and sort_metric.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8566'>#8566</a>] - An AutoML leaderboard extension is now available that includes model training time and model scoring time.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8558'>#8558</a>] - Exposed the location of Chunks in the REST API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8544'>#8544</a>] - Added a `rest_api_ping_timeout` option, which can stop a cluster if nothing has touched the REST API for the specified timeout.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8535'>#8535</a>] - Added support for Java 13.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8513'>#8513</a>] - H2O no longer performs an internal self-check when converting trees in H2O.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8840'>#8840</a>] - Fixed an XGBoost error on multinode with AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8818'>#8818</a>] - Added checkpointing to XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8664'>#8664</a>] - Users can now perform random grid search over target encoding hyperparameters
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8582'>#8582</a>] - Improved Grid Search testing in Flow.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11862'>#11862</a>] - When specifying a `stopping_metric`, H2O now supports lowercase and uppercase characters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9424'>#9424</a>] - Added a warning message to AutoML if the leaderboard is empty due to too little time for training.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9019'>#9019</a>] - In AutoML, blending frame details were added to event_log.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8879'>#8879</a>] - If early stopping is enabled, GBM can reset the ntree value. In these cases, added an `ntrees_actual` (Python)/`get_ntrees_actual` (R) method to provide the actual ntree value (whether CV is enabled or not) rather than the original ntree value set by the user before building a model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8808'>#8808</a>] - Refactored AutoML to improve integration with Target Encoding.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8708'>#8708</a>] - Exposed `get_automl` from `h2o.automl` in the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8701'>#8701</a>] - In GBM POJOs, one hot explicit EasyPredictModelWrapper now takes care of the encoding, and the user does not need to explicitly apply it.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8670'>#8670</a>] - Added support for numeric arrays to IcedHashMap.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8581'>#8581</a>] - Improved the AutoML Flow UI.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8574'>#8574</a>] - The `mae`, `rmsle`, and `aucpr` stopping metrics are now available in Grid Search.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8567'>#8567</a>] - When creating a hex.genmodel.easy.EasyPredictModelWrapper with contributions enabled, H2O now uses slf4j in the library, giving more control to users about when/where warnings will be printed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8491'>#8491</a>] - Moved the order of AUCPR in the list of values for `stopping_metric` to right after AUC.
</li>
</ul>
<h4>Engineering Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8541'>#8541</a>] - Removed unused code in UDPClientEvent.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8957'>#8957</a>] - Added examples to the Python Module documentation DRF chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8920'>#8920</a>] - Added examples to the Binomial Models section in the Python Python Module documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8905'>#8905</a>] - Added examples to the Multimonial Models section in the Python Python Module documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8903'>#8903</a>] - Added examples to the Clustering Methods section in the Python Module documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8902'>#8902</a>] - Added examples to the Regression section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8892'>#8892</a>] - Added examples to the Autoencoder section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8891'>#8891</a>] - Added examples to the Tree Class section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8872'>#8872</a>] - Added examples to the Assembly section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8867'>#8867</a>] - Added examples to the Node, Leaf Node, and Split Leaf Node sections in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8864'>#8864</a>] - Added examples to the H2O Module section in the Python documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8821'>#8821</a>] - Added examples to the H2OFrame section in the Python documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8804'>#8804</a>] - Documented support for `checkpointing` in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8802'>#8802</a>] - Added examples to the GroupBy section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8792'>#8792</a>] - Update to the supported platform table in the XGBoost chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8784'>#8784</a>] - Added R/Python examples to the metrics in Performance and Prediction section of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8782'>#8782</a>] - Added Parameter Appendix entries for CoxPH parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8744'>#8744</a>] - Added examples to the GBM section in the Python documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8728'>#8728</a>] - Added a new Reference entry to the Target Encoding documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8721'>#8721</a>] - Added examples to the KMeans section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8712'>#8712</a>] - Added examples to the CoxPH section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8697'>#8697</a>] - Added examples to the Deep Learning section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8667'>#8667</a>] - Added examples to the Stacked Ensembles section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8661'>#8661</a>] - Added new `use_spnego` option to the Starting H2O in R topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8654'>#8654</a>] - Added examples to the Target Encoding section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8652'>#8652</a>] - Added examples to the Aggregator section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8651'>#8651</a>] - Updated the XGBoost extramempercent FAQ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8636'>#8636</a>] - Added examples to the PCA section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8621'>#8621</a>] - Added a new section for Installing and Starting H2O in the Python Client documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8615'>#8615</a>] - Added examples to the SVD section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8604'>#8604</a>] - Improve the R and Python documentation for `search_criteria` in Grid Search.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8546'>#8546</a>] - Added an example using `predict_contributions` to the MOJO quick start.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8524'>#8524</a>] - Added examples to the PSVM section in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8512'>#8512</a>] - Added documentation for HGLM in the GLM chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8498'>#8498</a>] - Improved AutoML documentation:
<ul>
<li>aucpr is now an available stopping metric and sort metric for AutoML.</li>
<li>monotone_constraints can now be specified in AutoML.</li>
<li>Added modeling_plan option to list of AutoML parameters.</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8497'>#8497</a>] - MOJOs are now available for PCA.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8496'>#8496</a>] - MOJO models are now available for XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8494'>#8494</a>] - calibration_frame and calibrate_model are now available in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8493'>#8493</a>] - Added Java 13 to list of supported Java versions.
</li>
</ul>
Yau (3.26.0.11) - 12/05/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/11/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/11/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9051'>#9051</a>] - The Python client now fails with descriptive message when attempting to run on an unsupported Java version.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8740'>#8740</a>] - Fixed an issue that caused h2o to fail when running on Hadoop with `-internal_secure_connections`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8722'>#8722</a>] - H2OGenericEstimator can now be instantiated with no parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8692'>#8692</a>] - Multi-node H2O XGBoost now returns reproducible results.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8645'>#8645</a>] - Fixed the backend default values for the `inflection_point` and `smoothing` parameters in Target Encoder.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8634'>#8634</a>] - Users can now specify the `noise` parameter when running Target Encoding in the R client or in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8608'>#8608</a>] - MOJO reader now uses stderr instead of stdout to show warnings.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8584'>#8584</a>] - Fixed an issue that allowed SPNEGO athentication to pass with any HTTP-Basic header.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8578'>#8578</a>] - When connecting to H2O via the Python client, users can now specify `allowed_properties="cacert"`.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9406'>#9406</a>] - Added BroadcastJoinForTargetEncoding.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8669'>#8669</a>] - Introduced AllCategorical and Threshold TE application strategies.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8588'>#8588</a>] - Added a test to check XGBoost variable importance when trained on frames with shuffled input columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8587'>#8587</a>] - The package name for ai.h2o.org.eclipse.jetty.jaas.spi is now independent of the Jetty version.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8580'>#8580</a>] - The `offset_column` is now propogated to MOJO models.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8570'>#8570</a>] - Improved documentation for `stopping_metric` as it pertains to AutoML.
</li>
</ul>
Yau (3.26.0.10) - 11/7/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/10/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/10/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/41'>private-#41</a>] - Fixed an issue that caused H2O to ignore security configurations when running on Hadoop 3.x.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8614'>#8614</a>] - Added a `disable_flow` option that can be specified when starting H2O to disable access to H2O Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8599'>#8599</a>] - Version details are now exposed in cloud information.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8801'>#8801</a>] - Removed duplicate definition for sample_rate in DRF, as this is already defined in shared tree model parameters.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8601'>#8601</a>] - Fixed documentation for Logloss scorer.
</li>
</ul>
Yau (3.26.0.9) - 10/29/2019
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8803'>#8803</a>] - Fixed an issue that caused sort on a multinode cluster (for example, 2 nodes) to be much slower than a single node cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8702'>#8702</a>] - Fixed an issue that caused class conflicts between the released jars for h2o-genmodel-ext-xgboost and other Java packages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8683'>#8683</a>] - Export checkpoint no longer fails to export all models created during a grid search.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8630'>#8630</a>] - In the Python client, H2OFrame.drop no longer modifies parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8629'>#8629</a>] - Fixed an issue in the Python Client that caused model.actual_params to sometimes return a <property> object instead of a dict.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8663'>#8663</a>] - Fixed an issue that caused XGBoost to exhaust all memory on a node (-xmx+(1.2*-xmx)) on wide datasets.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8786'>#8786</a>] - Created a Technical Note (TN) describing how to use MOJO Import when importing models from a different H2O version. This TN is available here: <a href="https://github.com/h2oai/h2o-3/discussions/15523">https://github.com/h2oai/h2o-3/discussions/15523</a>.
</li>
</ul>
Yau (3.26.0.8) - 10/17/2019
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12867'>#12867</a>] - Fixed and ESPC row_layout assertion error that occurrend when run FrameTest.java.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8761'>#8761</a>] - In AutoML fixed an issue that resulted in poor predictions from SE on MNIST.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8742'>#8742</a>] - When saving files in Python, H2O now assumes the provided path is a directory even when an ending "/" is not included in the path.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8700'>#8700</a>] - The custom distribution function in GBM now works correctly for custom multinomial distributions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8695'>#8695</a>] - In Target Encoding, added blending of posterior and prior during imputation of unseen values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8688'>#8688</a>] - Removed H2O.STORE.clear() in FrameTest.java so that the deepSlice test could be enabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8681'>#8681</a>] - Python users can now specify `verify_ssl_certificates` and `cacert` when connecting to H2O.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8672'>#8672</a>] - Target Encoding now works correctly in Flow.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9136'>#9136</a>] - Base models can have different training_frames in blending mode in Stacked Ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8894'>#8894</a>] - Imported MOJO models now show parameters of the original model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8807'>#8807</a>] - Added the ability to clone ModelBuilder.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8735'>#8735</a>] - R client users can now specify the Boolean `use_spnego` parameter when starting H2O.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9358'>#9358</a>] - System level proxy is now bypassed when connecting to an H2O instance on localhost from R/Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8714'>#8714</a>] - Improved performance by sending data to and from external backend after a specified block size rather than by each item.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8689'>#8689</a>] - Disable HTTP TRACE requests.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8686'>#8686</a>] - Removed the "experimental" note in the AutoML chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8684'>#8684</a>] - Fixed a broken link in XGBoost documentation.
</li>
</ul>
Yau (3.26.0.6) - 10/1/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/6/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8769'>#8769</a>] - download_csv/download_all_logs now works correctly on HTTPS when using the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8765'>#8765</a>] - Fixed an error in PredictCSV unused column detection.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8739'>#8739</a>] - Fixed a potential deadlock issue with the AutoML leaderboard.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8732'>#8732</a>] - Model summary and model_performance output now display correctly in Zeppelin.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8731'>#8731</a>] - Added support for SPNEGO in h2odriver.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8729'>#8729</a>] - Fixed a missing chunk issue on the external backend.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8812'>#8812</a>] - In AutoML, you can now retrieve the leadernode using the REST API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8768'>#8768</a>] - Added support for MAPR 6.0 and 6.1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8743'>#8743</a>] - Added support for CDH 6.3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8737'>#8737</a>] - Added POJO support for one hot explicit encoding in GBM.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8752'>#8752</a>] - Added ability for countingErrorConsumer example to accumulate counters, not just for variables, but for each variable X and for each variable's value.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8854'>#8854</a>] - Added a new `melt` function. This is similar to Pandas `melt` and converts an H2OFrame to key-value representation while (optionally) skipping NA values. (This is the inverse operation to pivot.)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8771'>#8771</a>] - Added POJO support for XGBoost models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8755'>#8755</a>] - Removed the x-h2o-context-path header from H2O.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8750'>#8750</a>] - Upgraded H2O Flow to 0.10.7.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8838'>#8838</a>] - Moved MOJO Models topic from the Algorithms chapter in the User Guide to the Productionizing chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8773'>#8773</a>] - Added information about GPU usage for XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8741'>#8741</a>] - Added MapR 6.0 and 6.1 to list of supported Hadoop platforms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8725'>#8725</a>] - List all supported Java versions rather than saying Java 8 and greater.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8720'>#8720</a>] - Updated Deep Learning parameter descriptions for `rate`, `rate_annealing`, and `rate_decay`. Also added these to the Parameters Appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8719'>#8719</a>] - Updated the User Guide to indicate that POJOs are available for XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8711'>#8711</a>] - Added CDH 6.3 to list of supported Hadoop platforms.
</li>
</ul>
Yau (3.26.0.5) - 9/16/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8749'>#8749</a>] - Fixes a critical bug in Flow: Flow loads but user cannot perform any action.
</li>
</ul>
Yau (3.26.0.4) - 9/12/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9258'>#9258</a>] - Fixed several broken metric methods in the Python and R clients.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8856'>#8856</a>] - The temp folder is no longer deleted after running `h2o.mojo_predict_df`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8767'>#8767</a>] - Flow now works correctly in environments that already have a context path prefixed.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8858'>#8858</a>] - Introduced a Transform operation together with Target Encoding Model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8820'>#8820</a>] - Added support for CDH 5.15 and and CDH 5.16.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8839'>#8839</a>] - In the Flow > Models menu, moved MOJO Model to below the list of algorithms and re-labeled it "Import MOJO Model."
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8830'>#8830</a>] - Removed unnecessary read confirmation timeout.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8794'>#8794</a>] - Unified the Target Encoding API arguments with other models - using (x,y).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8785'>#8785</a>] - Target Encoder ignores non-categorical encoded columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8779'>#8779</a>] - Added "fetch mode" option to Flow. As a result, Hive users can now import tables from hive1.X from within Flow. Note that Hive 1.x doesn't support OFFSET. So for Hive 1.x, use import_hive_table or use non-distributed JDBC import (i.e., `fetch node = single`).
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8810'>#8810</a>] - Added `plug_values` to the Parameters Appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8783'>#8783</a>] - Added Python examples to CoxPH options.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8764'>#8764</a>] - Added Teradata to list of supported JDBC types.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8756'>#8756</a>] - Added CDH 5.15 and CDH 5.16 to list of supported Hadoop platforms.
</li>
</ul>
Yau (3.26.0.3) - 8/23/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12974'>#12974</a>] - Fixed an issue that caused an H2OResponseError after initialization of H2OSingularValueDecompositionEstimator.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9303'>#9303</a>] - AstGroup is no longer inconsistent after running it multiple times.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9213'>#9213</a>] - Fixed an issue that caused a mismatch between manual standard deviation and reported standard deviation for cross validation scores.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9027'>#9027</a>] - H2OFrame.split_frame() no longer leaks a _splitter object.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8913'>#8913</a>] - h2o.scale no longer modifies a frame in place. Instead, it now returns a new frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8901'>#8901</a>] - Users can export a model using java-rest-bindings.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8880'>#8880</a>] - Sorted grid search results by F2, F0point5 now correctly match the corresponding model metric.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8877'>#8877</a>] - Fixed an issue that caused XGBoost cox2 benchmark to fail with an NPE.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8863'>#8863</a>] - Tables with long titles now displaying properly for users who have installed Pandas.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8852'>#8852</a>] - ModelParametersSchemaV3 now displays the correct help messages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8849'>#8849</a>] - Fixed an issue that caused the MRTask to fail due to race-condition in creating a new Frame. Note that this issue only occurred when assertions were enabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8844'>#8844</a>] - Fixed an issue that caused GLM plots to fail in the Python client.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12359'>#12359</a>] - Added another mode to treat missing values: plug values. This value must be given by the user.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8828'>#8828</a>] - Implemented a re-try mechanism for requesting the flatfile on Hadoop.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9033'>#9033</a>] - Added Flow support for 2D partial plots.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8850'>#8850</a>] - In Flow, fixed issues with the NPM audit report.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9064'>#9064</a>] - Added support for predict_leaf_node_assignment() in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8914'>#8914</a>] - In Isolation Forest, improved documentation for aggregate depth and split ratios and described how these two values are calculated.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8908'>#8908</a>] - Removed MissingValuesHandling from XGBoost.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8929'>#8929</a>] - Added links to custom distribution and custom loss function demos.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8906'>#8906</a>] - Removed Java 7 from list of supported Java versions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8900'>#8900</a>] - Added Shapley example to predict_contributions documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8889'>#8889</a>] - Added H2ONode, H2OLeafNode, and H2OSplitNode to the Python client documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8883'>#8883</a>] - In XGBoost, removed "enum" from the list of available categorical_encoding options.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8873'>#8873</a>] - In GLM improved the documentation for handling of categorical values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8862'>#8862</a>] - Added predict_leaf_node_assignment to list of supported parameters in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8843'>#8843</a>] - The documentation for PSVM now indicates that it can be used for classification only.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8842'>#8842</a>] - The User Guide now includes all options that can be specified when running h2o.init() from the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8841'>#8841</a>] - Added bind_to_localhost to list of paramters for h2o.init() in the Python client docs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8817'>#8817</a>] - Updated GLM parameters. "PlugValues" can now be specified for missing_values_handling, and when specified, a new `plug_values` option is available.
</li>
</ul>
Yau (3.26.0.2) - 7/26/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12948'>#12948</a>] - Fixed an NPE error that occurred on models StackedEnsemble in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9043'>#9043</a>] - Improve the error message for rbind failures that resulted when rbinding datasets with long categorical levels.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8939'>#8939</a>] - In Flow, the scoring history deviance graph no longer displays if a custom distribution is not set.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8923'>#8923</a>] - pr_auc() now works correctly in the Python client.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9366'>#9366</a>] - Added support for Target Encoding MOJOs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9037'>#9037</a>] - Added support for Target Encoding transformation of data without a response column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8996'>#8996</a>] - Added TargetEncoderBuilder (estimator) and TargetEncoderModel (transformer).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8950'>#8950</a>] - Added detailed MOJO metrics for DRF, Isolation Forest, and GLM MOJO models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8948'>#8948</a>] - Added AUCPR to the list of available stopping_metric options.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9190'>#9190</a>] - In Flow, users can now upload a MOJO, and a generic model will automatically be created from it.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8951'>#8951</a>] - Removed duplicated code for obtaining logs in Java.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8947'>#8947</a>] - Improved error handling in the downloadLogs method.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8942'>#8942</a>] - Disabled autocomplete on the Flow login form.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8958'>#8958</a>] - Added an entry for upload_custom_metric in the Parameters Appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8954'>#8954</a>] - Added list of parameters that can be specified when building a Generic Model (MOJO import).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8945'>#8945</a>] - Updated documentation for MOJO Import.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8934'>#8934</a>] - Added "aucpr" to the list of available stopping_metric options.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8926'>#8926</a>] - Added an entry for export_checkpoints_dir in the Parameters Appendix.
</li>
</ul>
Yau (3.26.0.1) - 7/15/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yau/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yau/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12456'>#12456</a>] - Removed an unncessary warning in predict function that occcured when a test set was missing `fold_column`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9265'>#9265</a>] - AutoML no longer continues training models after a job cancellation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9173'>#9173</a>] - Fixed an issue that caused h2o Docker image builds to fail.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9078'>#9078</a>] - In XGBoost, parallel sparse matrix conversion is no longer using a non-threadsafe API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9062'>#9062</a>] - AutoML uses a default value of 5 for `score_tree_interval` with all algorithms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9055'>#9055</a>] - Fixed an issue that caused the Python client API to break when passing a frame to the constructor.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9029'>#9029</a>] - In Flow, you can now specify `blending_frrame` and `max_runtime_per_model` when running AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9050'>#9050</a>] - Frame Summary is now available when running the Python client in Zeppelin.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8974'>#8974</a>] - Fixed an issue that caused H2O.CLOUD._memary(idx).getTimestamp to return 0 rather than the timestamp of the remote node.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8970'>#8970</a>] - Fixed a link function NPE in MOJOs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8959'>#8959</a>] - Fixed the frame.tocsv signature. Instead of passing true, false, this now takes CSVStreamParams.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10966'>#10966</a>] - Added support for a custom Loss Metric in GBM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12929'>#12929</a>] - When running AutoML in R or Python, and EventLog is now available.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12930'>#12930</a>] - When polling an AutoML run, an EventLog displays now rather than a progress bar.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12945'>#12945</a>] - CoxPH is now available in the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12966'>#12966</a>] - Added support for SVM in the h2o-3 R and Python clients.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9138'>#9138</a>] - Added Isolation Forest to Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9120'>#9120</a>] - In XGBoost improved performance of moving sparse matrices to off-heap memory.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9112'>#9112</a>] - Logs from H2O can now be downloaded in plain text format.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12863'>#12863</a>] - Deprecated support for Java 7.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9020'>#9020</a>] - Fixed an issue that caused h2o.scale to corrupt the frame when run over a frame with categorical columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9012'>#9012</a>] - Removed the Deep Water booklet from H2O-3 builds.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12187'>#12187</a>] - AutoML runtime information is now stored and available in an EventLog.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12737'>#12737</a>] - Users can now pass an ID to training_frame in h2o.StackedEnsemble.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9216'>#9216</a>] - Added early stopping options to Isolation Forest.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9188'>#9188</a>] - Users can now build 2D Partial Dependence plots with the R and Python clients.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9147'>#9147</a>] - When loading MOJOs that were trained on older versions of H2O-3 into newer versions of H2O-3, users can now access all the information that was saved in the model object and use the MOJO to score.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9087'>#9087</a>] - Users can now specify a `row_index` parameter when building PDPs. This allows partial dependence to be calculated for a row.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9077'>#9077</a>] - Users can now specify a `row_index` parameter when building PDPs in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9058'>#9058</a>] - Enabled Java scoring for XGBoost MOJOs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9040'>#9040</a>] - User can now delete an AutoML instance and all its dependencies from any client (including models and other dependencies).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9014'>#9014</a>] - h2o.mojo_predict_csv() and h2o.mojo_predict_pandas() now accept a setInvNumNA parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9010'>#9010</a>] - Added support for TreeShap in DRF.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8985'>#8985</a>] - Added a `feature_frequencies` function in GBM, DRF, and IF, which retrieves the number of times a feature was used on a prediction path in a tree model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8984'>#8984</a>] - Users can now retrieve variable split information in the Isolation Forest output.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8990'>#8990</a>] - Created a SharedTreeMojoModelWithContributions class, which provides a central location of contribs for DRF and GBM MOJO.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8989'>#8989</a>] - ScoreContributionsTask is no longer abstract.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9174'>#9174</a>] - Clarified in the GLM docs that h2o-3 determines the values of alpha and theta by minimizing the negative log-likelihood plus the same Regularization Penalty.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9130'>#9130</a>] - Create initial, alpha version of SVM documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9076'>#9076</a>] - Added `upload_custom_distribution` to the Parameters Appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9026'>#9026</a>] - Removed note in XGBoost documentation indicating that "Multi-node support is currently available as a Beta feature."
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9023'>#9023</a>] - SVM R client documentation is now available.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9021'>#9021</a>] - Explained how the nthreads parameter can impact reproducibility.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9018'>#9018</a>] - Added stopping parameters to the Isolation Forest chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8994'>#8994</a>] - Fixed the parameters listing display for predict and predict_leaf_node_assignment in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8992'>#8992</a>] - DRF is now included in the list of supported algorithms for predict_contributions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8988'>#8988</a>] - Added more examples to the Predict topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8980'>#8980</a>] - Improved Data Manipulation Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8979'>#8979</a>] - Improved Modeling functions in the Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8977'>#8977</a>] - Improved the tree_class Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8976'>#8976</a>] - Improved the Model Metrics Python documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8975'>#8975</a>] - Improved GLM documentation by informing users that they can only specify a list in the GLM `interactions` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8971'>#8971</a>] - Updated Flow documentation to include Isolation Forest.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8968'>#8968</a>] - Improved the Python documentation for h2o.frame().
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/8967'>#8967</a>] - Added examples to the TargetEncoding Python documentation.
</li>
</ul>
Yates (3.24.0.5) - 6/18/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yates/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yates/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9237'>#9237</a>] - Fixed a segmentation fault that occurred when running XGBoost with `booster=gblinear`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9096'>#9096</a>] - Users can now rbind two frames when one frame contains all missing values in some of its columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9081'>#9081</a>] - ClearDKVTask now detects shared resources when deleting frames and models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9038'>#9038</a>] - Fixed a TypeError in Python debugging.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9115'>#9115</a>] - Fixed an issue that caused MOJO loading to fail when categorical values contained a newline character.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9100'>#9100</a>] - Users can now export a file directly to a compressed format (gzip) and choose a delimiter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9082'>#9082</a>] - Users can now specify which certificate alias to use when starting H2O with SSL.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9048'>#9048</a>] - Added Conda install instructions to the download page.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9039'>#9039</a>] - Users can now specify a custom separator for CSV export.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9169'>#9169</a>] - Fixed GLM std-error and Tweedie calculations.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9157'>#9157</a>] - Implemented dispersion factor optimization for Tweedie GLM.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9168'>#9168</a>] - The MOJO Tree Visualizer and Tree API no longer show categorical splits as numeric and string.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9122'>#9122</a>] - Improved the user experience with Target Encoding in R by providing more meaningful error messages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9110'>#9110</a>] - Users can now tokenize a frame to the Scala API to enable that using H2O's Word2Vec.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9105'>#9105</a>] - Defined several default values in the R API for Target Encoding.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9103'>#9103</a>] - Improved the user experience with Target Encoding in Python by providing more meaningful error messages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9101'>#9101</a>] - Set default values for blending hyperparameters in Target Encoding when using the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9097'>#9097</a>] - Fixed an issue that resulted in a "NaN undefined" label in the Flow cluster status.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9092'>#9092</a>] - Exposed ClearDKVTask via REST API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9083'>#9083</a>] - H2O-3 now provides a warning when using MOJO prediction with a test/validation dataset that has missing columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9056'>#9056</a>] - Upgraded the JTransforms library.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9232'>#9232</a>] - Added a Best Practices sub section to Starting H2O in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9156'>#9156</a>] - Added Target Encoding options to the Parameters appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9114'>#9114</a>] - Updated the description for the Tweedie family in the User Guide and in the GLM booklet.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9093'>#9093</a>] - Removed ologlog and oprobit from list of `link` options that can be specified in GLM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9063'>#9063</a>] - Upated documentation to indicate that predict_leaf_node_assignment is not supported with XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9034'>#9034</a>] - Added the new `-jks_alias` option to list of options that can be specified when starting H2O.
</li>
</ul>
Yates (3.24.0.4) - 5/28/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yates/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yates/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11194'>#11194</a>] - Fixed an error that occurred when applying as.matrix() to an h2o dataframe with numeric values of size ~ 600K x 300.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15445'>#15445</a>] - Introduced a new xgboost.predict.native.enable property, which ensures that H2OXGBoostEstimator will no longer always predicts the same value.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9186'>#9186</a>] - Users can now parse files from s3 using s3's directory URL with s3 protocol.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9154'>#9154</a>] - Fixed an issue that caused h2o.getModelTree to produce an "invalid object for slot nas" error when XGBoost produced a root-node only decision tree.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9153'>#9153</a>] - Improved performance of H2OXGBoost on OS X.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9150'>#9150</a>] - In Stacked Ensembles, fixed a categorical encoding mismatch error when building the ensemble. Users can now use SE on top of base models that are trained with categorical encoding.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9146'>#9146</a>] - In Isolation Forest, you can now specify that mtries = the number of features.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9142'>#9142</a>] - Fixed an issue that caused XGBoost to produce a tree with split features being all NA.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9141'>#9141</a>] - In h2o.getModelTree, when retrieving a threshold for values that are all NAs, updated the description to state that the "Split value is NA."
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9140'>#9140</a>] - Fixed an issue that caused trivial features with NAs to be given inflated importance when monotonicity constraints was enabled. As a result, variable importance values were incorrect.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9139'>#9139</a>] - Fixed an NPE issue at water.init.HostnameGuesser when trying to launch a Sparkling Water cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9134'>#9134</a>] - Removed internal_cv_weights from h2o.predict_contributions() output when the prediction was used on a fold column from a model run with nfolds.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9109'>#9109</a>] - Models that use Label Encoding no longer predict incorrectly on test data.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9107'>#9107</a>] - Predictions now work correctly on a subset of training features when using categorical_encoding.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9098'>#9098</a>] - Fixed an issue that caused XGBoost to format non-integer numbers (doubles, floats) using Locale.ENGLISH to ensure that a decimal point "." was used instead of a comma ",".
This locale setting grouped large numbers by thousands and split the groups with ",", which was unparseable to XGBoost.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9151'>#9151</a>] - Added support for CDH 6.2.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9127'>#9127</a>] - Users can now specify an external IP for h2odriver callback.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9111'>#9111</a>] - Added a "toCategoricalCol" helper function for column type conversion.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9108'>#9108</a>] - Renamed "Generic Models" to "MOJO Import" in the documentation.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9144'>#9144</a>] - Added CDH 6.2 to list of supported Hadoop platforms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9119'>#9119</a>] - Added the import_hive_table() and import_mojo() functions to the R HTML documentation.
</li>
</ul>
Yates (3.24.0.3) - 5/7/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yates/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yates/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12817'>#12817</a>] - Updated H2O-3 Plotting Functionality to be Compatible with Matplotlib Version 3.0.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9240'>#9240</a>] - Flow now shows the correct long value of a seed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9230'>#9230</a>] - Fixed an issue that cause Rapids string operations on enum (categorical) columns to yield counterintuitive results.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9223'>#9223</a>] - Fixed an issue that caused monotonicity constraint in XGBoost to fail with certain parameters
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9218'>#9218</a>] - Fixed an ArrayIndexOutOfBounds error. that occurred when parsing quotes in CSV files.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9210'>#9210</a>] - Fixed an error with Grid Search that caused the API to print errors not related to model CURRENTLY being added to the grid, but for all previous failures. This occurred even when the model was not added to the grid due to failure.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9195'>#9195</a>] - Fixed an exception that occurred when requesting Jobs from h2o.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9187'>#9187</a>] - When using Python 2.7, fixed an issue with non-ascii character handling in the as_data_frame() method.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9177'>#9177</a>] - Predicting on a dataset that has a response column with domain in a different order no longer leads to memory leaks.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9175'>#9175</a>] - Fixed an issue with retrieving details of a GLM model in Flow due to lack of support for long seeds.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9207'>#9207</a>] - Simplified the directory structure of logs within downloaded zip archives.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9198'>#9198</a>] - Upgrades XGBoost to latest stable build.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9191'>#9191</a>] - Users can how import and upload MOJOs in R and Python using `import_mojo()` and `upload_mojo()`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9176'>#9176</a>] - It is now possible to retrieve a list of features from a trained model.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12872'>#12872</a>] - Enhanced the GBM Reproducibility FAQ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9170'>#9170</a>] - Added information about the Target Encoding smoothing parameter to the User Guide.
</li>
</ul>
Yates (3.24.0.2) - 4/16/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yates/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yates/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9399'>#9399</a>] - In the R client, fixed a caching issue that caused tests to fail when running commands line by line after running the entire test at once.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9255'>#9255</a>] - Fixed an issue that caused the h2o.upload_custom_metric to fail when using python3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9254'>#9254</a>] - Fixed an issue that caused h2o.upload_custom_metric to fail on data that includes strings.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9253'>#9253</a>] - Fixed an issue with the K-Means_Example.flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9252'>#9252</a>] - The IP:port that is shown for logging now matches the IP:port that is described in the makeup of the cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9247'>#9247</a>] - In XGBoost, fixed an AIOOB issue that occurred when running large data.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9234'>#9234</a>] - H2O-hive is now published to Maven central.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9231'>#9231</a>] - The Rapids as.factor operation no longer automatically converts non-ASCII strings to sanitized forms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9229'>#9229</a>] - Fixed an AIOOB error in the AUC builder.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9225'>#9225</a>] - AUCBuilder now finds the first bin to merge when merging per-chunk histograms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9217'>#9217</a>] - When running H2O on Hadoop, Hadoop now writes only to its container directory.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9208'>#9208</a>] - Users now receive a warning if two different versions of H2O are trying to communicate on the same node.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9205'>#9205</a>] - Fixed an issue that caused the H2O Python package to fail to load on a fresh install from pip.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9193'>#9193</a>] - Fixed an error that occurred when running multiple concurrent Group-By operations.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9312'>#9312</a>] - The new GCP Marketplace offering contains the option to add a network tags script.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12887'>#12887</a>] - Added Python examples to the Target Encoding topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9224'>#9224</a>] - Fixed links to Sparkling Water topics in the Sparkling Water FAQ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9201'>#9201</a>] - In CoxPH chapter, changed the link for the available R demo.
</li>
</ul>
Yates (3.24.0.1) - 3/31/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-yates/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-yates/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12988'>#12988</a>] - The AutoMLTest.java test suite now runs correctly on a local machine.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9430'>#9430</a>] - Fixed an issue in as_date that occurred when the column included NAs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9411'>#9411</a>] - AutoML no longer fails if one of the Stacked Ensemble models is deleted.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9391'>#9391</a>] - Removed elipses after the H2O server link when launching the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9390'>#9390</a>] - In Deep Learning, fixed an issue that occurred when running one-hot-encoding on categoricals.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9359'>#9359</a>] - When running GBM in R without specifically setting a seed, users can now extract the seed that was used to build the model and reproduce that model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9355'>#9355</a>] - In predictions, fixed an issue that resulted in a "Categorical value out of bounds error" when calling a model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9337'>#9337</a>] - The Python API no longer reverses the labels for positive and negative values in the standardized coefficients plot legend.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9278'>#9278</a>] - In R, fixed an issue that cause group_by mean to only calculate one column when multiple columns were specified.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9274'>#9274</a>] - Fixed an issue that caused the confusion_matrix method to return matrices for other metrics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9267'>#9267</a>] - Fixed an issue that resulted in a "Categorical value out of bounds error" when calling a model using Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9264'>#9264</a>] - Improved the error message that displays when a user attempts to modify an Enum/categorical column as if it were a string.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9257'>#9257</a>] - Rows that start with a symbol are no longer dropped during the import process.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9256'>#9256</a>] - Fixed an SVM import failure.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9248'>#9248</a>] - Fixed an issue that caused the default StackedEnsemble prediction to fail when applied to a test dataset without a response column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9245'>#9245</a>] - Fixed handling of BAD state in CategoricalWrapperVec.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11559'>#11559</a>] - Added Blending mode to Stacked Ensembles, which can be specified with the `blending_frame` parameter. With Blending mode, you do not use cross-validation preds to train the metalearner. Instead you score the base models on a holdout set and use those predicted values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12655'>#12655</a>] - Model output now includes column names and types.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12663'>#12663</a>] - AutoML now includes a max_runtime_secs_per_model option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15504'>#15504</a>] - In GLM, added support for negative binomial family.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12828'>#12828</a>] - ExposeD Java target encoding to R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12902'>#12902</a>] - For GBM and XGBoost models, users can now generate feature contributions (SHAP values).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12968'>#12968</a>] - Added support for Generic Models, which provide a means to use external, pretrained MOJO models in H2O for scoring. Currently only GBM, DRF, IF, and GLM MOJO models are supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9439'>#9439</a>] - Added the blending_frame parameter to Stacked Ensembles in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9423'>#9423</a>] - Added an include_algos parameter to AutoML in the R and Python APIs. Note that in Flow, users can specify exclude_algos only.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9284'>#9284</a>] - In the R and Python clients, added a function that calculates the chunk size based on raw size of the data, number of CPU cores, and number of nodes.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9280'>#9280</a>] - Added ability to import from Hive using metadata from Metastore.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9266'>#9266</a>] - Users can now choose the database where import_sql_select creates a temporary table.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9259'>#9259</a>] - Added support for monotonicity constraints for binomial GBMs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9250'>#9250</a>] - Users can now define custom HTTP headers using an `-add_http_header` option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9238'>#9238</a>] - XGBoost MOJO now uses Java predictor by default.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11858'>#11858</a>] - Fixed an issue that caused the pyunit_lending_club_munging_assembly_large.py and pyunit_assembly_munge_large.py tests to sometimes fail when run inside a Docker container.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12728'>#12728</a>] - Simplified and improved the GLM COD implementation.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12357'>#12357</a>] - SQLite support is available via any JDBC driver in streaming mode.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12841'>#12841</a>] - Updated Retrofit and okHttp dependecies.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12961'>#12961</a>] - Target Encoding is now available in the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9443'>#9443</a>] - Moved StackedEnsembleModel to hex.ensemble packages. In prior versions, this was in a root hex package.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9431'>#9431</a>] - Secret key ID and secret key are available for s3:// AWS protocol.
<ul><li>This can be done in the R client using:
<br>h2o.setS3Credentials(accessKeyId, accesSecretKey) </li>
<br><li>and in the Python client using:
<br>from h2o.persist import set_s3_credentials
<br>set_s3_credentials(access_key_id, secret_access_key)</li></ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9403'>#9403</a>] - Users can now specify AWS credentials at runtime.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9367'>#9367</a>] - The new blending_frame parameter is now available in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9289'>#9289</a>] - Fixed an error in the Javadoc for the Frame.java sort function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9261'>#9261</a>] - Fixed Hive delegation token generation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9236'>#9236</a>] - Reordered the algorithms train in AutoML and prioritized hardcoded XGBoost models.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11853'>#11853</a>] - Removed FAQ indicating that Java 9 was not yet supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12968'>#12968</a>] - Added a "Generic Models" chapter to the Algorithms section.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9440'>#9440</a>] - Added the blending_frame parameter to Stacked Ensembles documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9341'>#9341</a>] - Added information about the Negative Binomial family to the GLM booklet and the user guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9331'>PUBDV-6289</a>] - Improved the R and Python client documentation for the `sum` function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9292'>#9292</a>] - Added include_algos,e xclude_algos, max_models, and max_runtime_secs_per_model examples to the Parameters appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9262'>#9262</a>] - In the User Guide and R an Python documentation, replaced references to "H2O Cloud" with "H2O Cluster".
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9249'>#9249</a>] - Added information about predict_contributions to the Performance and Prediction chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9243'>#9243</a>] - In the GBM chapter, noted that monotone_constraints is available for Bernoulli distributions in addition to Gaussian distributions.
</li>
<li>Improved the GBM Reproducibility FAQ.</li>
</ul>
Xu (3.22.1.6) - 3/13/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xu/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xu/6/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9288'>#9288</a>] - In GBM, added a check to ensure that monotonicity constraints can only be used when distribution="gaussian".
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9282'>#9282</a>] - Fixed an issue that caused decreasing monotonic constraints to fail to work correctly. Min-Max bounds are now properly propagated to the subtrees.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9281'>#9281</a>] - Added internal validation of monotonicity of GBM trees.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9286'>#9286</a>] - Updated the description of monotone_constraints for GBM. This option can only be used for gaussian distributions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9277'>#9277</a>] - Improved documentation for the EC2 and S3 storage topic for AWS Standalone instances (http://docs.h2o.ai/h2o/latest-stable/h2o-docs/cloud-integration/ec2-and-s3.html#aws-standalone-instance).
</li>
</ul>
Xu (3.22.1.5) - 3/4/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xu/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xu/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9338'>#9338</a>] - Fixed an issue that caused stratified_split to fail when run on same column twice.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9330'>#9330</a>] - Fixed an error that occurred when retreiving AutoML leader model with max_models = 1 in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9328'>#9328</a>] - Fixed an issue that ersulted in an extra NA row in the GLM variable importance frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9322'>#9322</a>] - h2odriver now works correctly on MapR.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9320'>#9320</a>] - Flow no longer displays an error when searching for a file without first providing a path.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9317'>#9317</a>] - GBM monotonicity constraints now correctly preserves the exact monotonicity.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9332'>#9332</a>] - Fixed the warning message that displays for categorical data with more then 10,000,000 values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9316'>#9316</a>] - Users can now download logs from R after connecting via Steam.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9309'>#9309</a>] - In AutoML, created new partition rules for generating new validation and leaderboard frames when cross validation is disabled and validation/leaderboard frames are not provided:
<ul>
<li>If only the validation frame is missing: training/validation = 90/10.
</li>
<li>If only the leaderboard frame is missing: training/leaderboard = 90/10.
</li>
<li>If both the validation and leaderboard frames are missing: training/validation/leaderboard = 80/10/10.
</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9301'>#9301</a>] - Fixed resolution of `spark-shell --packages "ai.h2o:h2o-algos:<vesion>"` by Spark Ivy resolver.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9290'>#9290</a>] - Fixed an issue that caused h2o driver to fail to start when Hive was not configured.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9350'>#9350</a>] - In Isolation Forest, fixed an issue that caused the minimum and maximum path length to not be correctly calculated when there are no OOB observations.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9326'>#9326</a>] - A `check_constant_response` option is available in DRF and GBM. When enabled (default), then an exception is thrown if the response column is a constant value.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12417'>#12417</a>] - When running XGBoost on Hadoop, recommend that users set -extramempercent to 120.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9334'>#9334</a>] - Added the new check_constant_response option to the GBM and DRF chapters. Also added an example usage to the Parameters Appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9319'>#9319</a>] - Added a description of the AUCPR metric to the Model Performance section in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9308'>#9308</a>] - Fixed the Random Grid Search in Python example in the Grid Search chapter.
</li>
</ul>
Xu (3.22.1.4) - 2/15/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xu/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xu/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9379'>#9379</a>] - Users can now save and load Isolation Forest models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9357'>#9357</a>] - In K-Means, fixed and issue in which time columns were treated as if they were categorical.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9354'>#9354</a>] - Fixed Autoencoder `calculateReconstructionErrorPerRowData` error and set the default value of the result MSE to -1.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/47'>private-#47</a>] - When using h2o.import_sql_table to read from a Hive table, the username and password no longer appear in the logs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9412'>#9412</a>] - Monotone constraints are now exposed in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9344'>#9344</a>] - The check for constants in response columns is now optional for all models.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12879'>#12879</a>] - Added to the documentation that MOJO/POJO predict cannot parse columns enclosed in double quotes (for example, ""2"").
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9445'>#9445</a>] - Updated the description for Gini in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9436'>#9436</a>] - Fixed the equation for Tweedie Deviance in the GLM booklet and in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9420'>#9420</a>] - Added a "Tokenize Strings" topic to the Data Manipulation chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9376'>#9376</a>] - Added `predict_leaf_node_assignment` information to the User Guide in the Performance and Prediction chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9368'>#9368</a>] - Noted in the documentation that the `custom` and `custom_increasing` stopping metric options are not available in the R client.
</li>
</ul>
Xu (3.22.1.3) - 1/25/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xu/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xu/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9433'>#9433</a>] - Improved error handling for a wrong Hive JDBC connector error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9388'>#9388</a>] - Fixed an issue that caused H2O clusters to fail to come up on Cloudera 6 with HTTPS.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9404'>#9404</a>] - Added Hive with Kerberos support for H2O on Hadoop.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9401'>#9401</a>] - Updated the default value for min_rows in the User Guide when used with XGBoost, DRF, and Isolation Forest.
</li>
</ul>
Xu (3.22.1.2) - 1/18/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xu/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xu/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15450'>#15450</a>] - In Flow, fixed an issue that caused POJOs, MOJOs, and genmodel.jar to fail to download. This occurred when Flow was launched via Enterprise Steam and in any deployment where user_context was specified.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9442'>#9442</a>] - Fixed an issue that caused H2OTree to fail with Isolation Forest models trained on data with categorical columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9441'>#9441</a>] - When a new tree is assembled from a model, the root node now includes information about the split feature in the description array.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9438'>#9438</a>] - Fixed an issue where Flow failed to provide the ability to ignore certain columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9427'>#9427</a>] - In Flow, fixed an issue where users were not able to select a frame when splitting a dataset.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9422'>#9422</a>] - Setting the `ignored_columns` parameter via the Python API now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9421'>#9421</a>] - Fixed an issue that caused H2O to hang in Sparkling Water deployments.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9419'>#9419</a>] - Splitting frames now works correctly in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9418'>#9418</a>] - Import SQL Table now works correctly in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9416'>#9416</a>] - Fixed an issue with imports in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9415'>#9415</a>] - Fixed interaction pairs for GLM in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9413'>#9413</a>] - Fixed broken "Combine predictions with frame" in Flow.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12977'>#12977</a>] - Added support for HDP 3.1.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9448'>#9448</a>] - Fixed the pyunit_pubdev_3500_max_k_large.py unit test.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9447'>#9447</a>] - Fixed the runit_PUBDEV_5705_drop_columns_parser_gz.R unit test.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9452'>#9452</a>] - Increased the XGBoost stress test timeout.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9431'>#9431</a>] - Implemented secret key credentials for s3:// AWS protocol.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9414'>#9414</a>] - Renamed .jade files to .pug.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9454'>#9454</a>] - Added HDP 3.0 and 3.1 to list of supported Hadoop versions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9429'>#9429</a>] - Updated wording for Kmeans Scoring History Graph. This graph shows the number of iterations vs. within the cluster’s sum of squares.
</li>
</ul>
Xu (3.22.1.1) - 12/28/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xu/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xu/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12108'>#12108</a>] - PCA tests now work correctly with the "from h2o.estimators.pca import H2OPrincipalComponentAnalysisEstimator" import statement.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12805'>#12805</a>] - Fixed an AutoMLTest test that was leaking keys in KeepCrossValidationFoldAssignment test.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12923'>#12923</a>] - Reduced the Invocation JMH level setup/teardown to only the training model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12956'>#12956</a>] - In XGBoost, the default value of L2 regularization for tree models is now 1, which is consistent with native XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12986'>#12986</a>] - Fixed an issue that caused Stacked Ensembles to fail with GLM metalearner when the same H2O instance was used to train a GLM multinomial classification model with more classes than what is used in Stacked Ensembles.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12133'>#12133</a>] - Users can now specify `custom` and `custom_increasing` when setting the `stopping_criteria` parameter in GBM and DRF.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12625'>#12625</a>] - Checkpoints can now be exported when running Grid Search or AutomL.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12746'>#12746</a>] - Added support for CDH 6.0, which includes Hadoop 3 support. Be sure to review <a href="https://www.cloudera.com/documentation/enterprise/6/release-notes/topics/rg_cdh_600_release_notes.html">https://www.cloudera.com/documentation/enterprise/6/release-notes/topics/rg_cdh_600_release_notes.html</a> for more information.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12802'>#12802</a>] - Fixed an AutoMLTest that was leaking keys.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12926'>#12926</a>] - Added a test that runs multiple `nfolds>0` DRF models in parallel.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12983'>#12983</a>] - Added support for CDH 6.1
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12674'>#12674</a>] - Hadoop builds now work with Jetty 8 and 9.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12749'>#12749</a>] - R examples in the R package docs now use Hadley's style guide.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12894'>#12894</a>] - Added documentation for the new stopping_metric options in GBM and DRF.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12984'>#12984</a>] - Added CDH 6 and 6.1 to list of supported Hadoop versions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12985'>#12985</a>] - In the XGBoost chapter, updated the default value for reg_lambda to be 1.
</li>
</ul>
Xia (3.22.0.5) - 1/16/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xia/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xia/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9421'>#9421</a>] - Fixed an H2O hang issue in Sparkling Water deployments.
</li>
</ul>
Xia (3.22.0.4) - 1/4/2019
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xia/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xia/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15450'>#15450</a>] - In Flow, fixed an issue that caused POJOs, MOJOs, and genmodel.jar to fail to download. This occurred when Flow was launched via Enterprise Steam and in any deployment where user_context was specified.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9453'>#9453</a>] - On the external backedn, H2O now explicitly passes the timestamp from the Spark Driver node.
</li>
</ul>
Xia (3.22.0.3) - 12/21/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xia/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xia/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12682'>#12682</a>] - Fixed an issue with the REST API. Calling "get model" no longer returns 0 for the timestamp of the model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12808'>#12808</a>] - The PySparking client no longer hangs after re-connecting to the H2O external backend.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12838'>#12838</a>] - Fixed an OOM issue in h2o.arrange.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12905'>#12905</a>] - Fixed an issue that caused importing Pargue files with large Double data to fail.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12918'>#12918</a>] - After applying group_by to a time stamped column, the original time stamp format is now retained.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12921'>#12921</a>] - In AutoML, cross-validation metrics are now used for early stopping by default. Because of this, the validation_frame argument is now ignored unless nfolds==0 and, in that case, will be used for early stopping.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12936'>#12936</a>] - Fixed an issue that caused the MOJO visualizer to fail for Isolation Forest models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12939'>#12939</a>] - StackedEnsembleMojoModel is now serializable.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15449'>#15449</a>] - In the R client, fixed an error that occurrred when running getModelTree.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15450'>#15450</a>] - In Flow, fixed an issue that caused POJOs, MOJOs, and genmodel.jar to fail to download. This occurred when Flow was launched via Enterprise Steam and in any deployment where user_context was specified.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12947'>#12947</a>] - Fixed the formula used for calculating L2 distance.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12952'>#12952</a>] - The Python client now allows users to enable XGBoost compare with any H2O frame. The convert_H2OFrame_2_DMatrix method accepts any H2O frame and can convert it to valid data for native XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12954'>#12954</a>] - H2O XGBoost now reports correct variable importances. The variable importances are computed from the gains of their respective loss functions during tree construction.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12955'>#12955</a>] - Users can now save PDP plots.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15454'>#15454</a>] - Fixed an issue that resulted in a SQL exception when connecting H2O to a SQL server and importing a table.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12969'>#12969</a>] - Fixed an issue with GCS support on Hadoop environments.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14868'>#14868</a>] - Added monotonic variables for GBM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12877'>#12877</a>] - EasyPredictModelWrapper now calculates reconstruction errors for AutoEncoder.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12931'>#12931</a>] - When running a grid search, a timesteamp column was added that shows when each model was added to the grid summary table.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12717'>#12717</a>] - In GBM, users can now specify the `monotone_constraints` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12944'>#12944</a>] - Prediction contributions from each tree from MOJO to easywrapper are now exposed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12946'>#12946</a>] - Updated Gradle to version 5.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12950'>#12950</a>] - Fixed the output of rankTsv in the AutoML leaderboard.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11266'>#11266</a>] - Updated the Prediction section to include information on how the prediction threshold is selected for classification problems.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12943'>#12943</a>] - Updated the description of enum_limited to indicate that T=1024.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12978'>#12978</a>] - In the GBM chapter, added `monotone_constraints` to list of available parameters.
</li>
</ul>
Xia (3.22.0.2) - 11/21/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xia/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xia/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10196'>#10196</a>] - Fixed an issue that caused ARFF parser to parse some file incorrectly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11616'>#11616</a>] - When performing a grid search in Python, fixed an issue that caused all models to return a model.type of "supervised."
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12220'>#12220</a>] - When running DRF in the Python client, checkpointing on new data now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12721'>#12721</a>] - Fixed an issue that caused the confusion matrix recall and precision values to be switched.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12883'>#12883</a>] - In the Python client, fixed an issue that caused the `offset_column` parameter to be ignored when it was passed in the GLM train statement.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12889'>#12889</a>] - The H2O Tree Handler now works correctly on Isolation Forest models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12893'>#12893</a>] - When running AutoML, fixed an issue that resulted in a "Failed to get metric: auc from ModelMetrics type BinomialGLM" message.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12896'>#12896</a>] - In Flow, Precision and Recall definitions are no longer inverted in the confusion matrix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12898'>#12898</a>] - Fixed the error message that displays when converting from a pandas dataframe to an h2oframe in Python 3.6.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12900'>#12900</a>] - In XGBoost, fixed an issue that resulted in a "Maximum amount of file descriptors hit" message.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12906'>#12906</a>] - Fixed the description of sample_rate in Isolation Forest.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12909'>#12909</a>] - Cross validation models are no longer deleted by default.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12911'>#12911</a>] - When viewing an AutoML leaderboard, fixed an issue that resulted in an ArrayIndexOutOfBoundsException if `sort_metric` was specified but no model was built.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12621'>#12621</a>] - Added monotonicity constraints to H2O XGBoost.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12886'>#12886</a>] - When generating MOJOs, h2o-genmodel.jar now includes a check for MOJO version 1.3 to determine whether the ho2-genmodel.jar and the MOJO version can work together. Prior versions of h2o-3 did not include MOJO 1.3, and as a result, MOJOs silently returned predicted values executed on an empty vector.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12561'>#12561</a>] - With a new `skipped_columns` option, users can now specify to drop specific columns before parsing. Note that this functionality is not supported for SVMLight or Avro file formats.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12908'>#12908</a>] - The GLM multinomial coefficient table now includes the original levels as column names.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10134'>#10134</a>] - Created new Performance & Prediction and Variable Importance sections in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12184'>#12184</a>] - Updatd the default value of `categorical_encoding` for XGBoost. This defaults to Auto (which is one_hot_encoding).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12860'>#12860</a>] - In the parameter entry for `weights_column`, updated the example to exclude the weight column in the list of predictors.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12864'>#12864</a>] - In the DRF FAQ, updated the "What happens when you try to predict on a categorical level not seen during training?" question.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12873'>#12873</a>] - TargetingEncoder is now included in the Python module docs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12888'>#12888</a>] - In GLM, updated the documentation to indicate that coordinate_descent is no longer experimental.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12910'>#12910</a>] - Added default values for `max_depth`, `sample_size`, and `sample_rate`. Also added a parameter description entry for `sample_size`, showing an Isolation Forest example.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12927'>#12927</a>] - Added the new `monotone_constraints` option to the XGBoost chapter.
</li>
</ul>
Xia (3.22.0.1) - 10/26/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-xia/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-xia/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11898'>#11898</a>] - In Python, the metalearner method is only available for Stacked Ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12517'>#12517</a>] - Fixed an issue that caused micro benchmark tests to fail to run in the jmh directory.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12522'>#12522</a>] - Fixed an issue that caused H2O to fail to export dataframes to S3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12600'>#12600</a>] - Added the `keep_cross_validation_models` argument to Grid Search.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12601'>#12601</a>] - Improved efficiency of the `keep_cross_validation_models` parameter in AutoML
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12632'>#12632</a>] - Simplified the comparison of H2OXGBoost with native XGBoost when using the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15502'>#15502</a>] - Fixed JDBC ingestion for Teradata databases.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12677'>#12677</a>] - In the Python client and the Java API, multiple runs of the same AutoML instance no longer fail training new "Best Of Family" SE models that would include the newly generated models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12725'>#12725</a>] - Fixed an issue that resulted in an AssertionError when calling `cbind` from the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12733'>#12733</a>] - AutoML now enforces case for the `sort_metric` option when using the Java API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12755'>#12755</a>] - In AutoML, StackEnsemble models are now always trained, even if we reached `max_runtime_secs` limit.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12756'>#12756</a>] - In the R client, added documentation for helper functions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12774'>#12774</a>] - Renamed `x` to `X` in the H2O-sklearn fit method to be consistent with the sklearn API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12776'>#12776</a>] - Merging datasets now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12782'>#12782</a>] - Building on Maven with h2o-ext-xgboost on versions later than 3.18.0.11 no longer results in a dependency error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12784'>#12784</a>] - Fixed a Java 11 ORC file parsing failure.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12803'>#12803</a>] - Upgraded the version of the lodash package used in H2O Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12815'>#12815</a>] - `-ip localhost` now works correctly on WSL.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12819'>#12819</a>] - CSV/ARFF Parser no longer treats blank lines as data lines with NAs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12824'>#12824</a>] - Starting h2o-3 from the Python Client no longer fails on Java 10.0.2.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12843'>#12843</a>] - Fixed an issue that caused StackedEnsemble MOJO model to return an "IllegalArgumentException: categorical value out of range" message.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12844'>#12844</a>] - Removed the "nclasses" parameter from tree traversal routines.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12846'>#12846</a>] - Exposed H2OXGBoost parameters used to train a model to the Python API. Previously, this information was visible in the Java backend but was not passed back to the Python API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12847'>#12847</a>] - Removed "illegal reflective access" warnings when starting H2O-3 with Java 10.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12852'>#12852</a>] - In Stacked Ensembles, changes made to data during scoring now apply to all models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12853'>#12853</a>] - When running AutoML in Flow, updated the list of algorithms that can ber selected in the "Exclude These Algorithms" section.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12042'>#12042</a>] - Individual predictions of GBM trees are now exposed in the MOJO API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12245'>#12245</a>] - Exposed target encoding in the Java API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12266'>#12266</a>] - The `keep_cross_validation_fold_assignment` option is now available in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12470'>#12470</a>] - Added support for the Isolation Forest algorithm in H2O-3. Note that this is a Beta version of the algorithm.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12527'>#12527</a>] - Added the `keep_cross_validation_fold_assignment` option to AutoML in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12540'>#12540</a>] - `h2o.connect` no longer ignores `strict_version_check=FALSE` when connecting to a Steam cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12551'>#12551</a>] - Created an R demo for CoxPH. This is available <a href="https://github.com/h2oai/h2o-3/blob/master/h2o-r/demos/rdemo.word2vec.craigslistjobtitles.R">here</a>.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12630'>#12630</a>] - It is now possible to combine two models into one MOJO, with the second model using the prediction from the first model as a feature. These models can be from any algorithm or combination of algorithms except Word2Vec.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12704'>#12704</a>] - Implemented h2oframe.fillna(method='backward').
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12825'>#12825</a>] - Improved speed-up of AutoML training on smaller datesets in client mode (Sparkling Water).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12827'>#12827</a>] - Exposed Java Target Encoding in the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12836'>#12836</a>] - Users can now specify a `-features` parameter when starting h2o from the command line. This allows users to remove experimental or beta algorithms when starting H2O-3. Available options for this parameter include `beta`, `stable`, and `experimental`.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11391'>#11391</a>] - Added XGBoost to AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12552'>#12552</a>] - Added an option to allow users to use a user-specified JDBC driver.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12578'>#12578</a>] - Exposed `pr_auc` to areas where you can find AUC, including scoring_history, model summary. Also added h2o.pr_auc() in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12753'>#12753</a>] - Added support for Java 11.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12849'>#12849</a>] - Improved the AutoML documentation in the User Guide.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12451'>#12451</a>] - Added a `MAX_USR_CONNECTIONS_KEY` argument to limit number of sessions for import_sql_table.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12528'>#12528</a>] - Improved performance gap when importing data using Hive2.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12575'>#12575</a>] - Improved and cleaned up output for the h2o.mojo_predict_csv and h2o.mojo_predict_df functions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12598'>#12598</a>] - Users can now visualize XGBoost trees when running predictions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12616'>#12616</a>] - Added weights to partial depenced plots. Also added a level for missing values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12675'>#12675</a>] - Users can now download the genmodel.jar in Flow for completed models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12738'>#12738</a>] - In AutoML, changed the default for `keep_cross_validation_models` and `keep_cross_validation_predictions` from True to False.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12740'>#12740</a>] - Added support for predicting using the XGBoost Predictor.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12761'>#12761</a>] - In XGBoost, optimized the matrix exchange between Java and native C++ code.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12765'>#12765</a>] - Improved the h2o-3 README for installing in R and IntelliJ IDEA.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12778'>#12778</a>] - Introduced a simple "streaming" mode that allows H2O to read from a table using basic SQL:92 constructs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12780'>#12780</a>] - In AutoML, `stopping_metric` is now based on `sort_metric`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12801'>#12801</a>] - The requirements.txt file now includes the Colorama version.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12809'>#12809</a>] - In lockable.java, delete is now final in order to prevent inconsistent overrides.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12812'>#12812</a>] - Reverted AutoML naming change from Auto.Algo to Auto.algo.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12848'>#12848</a>] - In AutoML, automatic partitioning of the valiation frame now uses 10% of the training data instead of 20%.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12850'>#12850</a>] - Changed model and grid indexing in autogenerated model names in AutoML to be 1 instead of 0 indexed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12865'>#12865</a>] - Allow public access to H2O instances started from R/Python. This can be done with the new `bind_to_localhost` (Boolean) parameter, which can be specified in `h2o.init()`.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11389'>#11389</a>] - Added Scala and Java examples to the Building and Extracting a MOJO topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11472'>#11472</a>] - Added a Scala example to the Stacked Ensembles topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12798'>#12798</a>] - Added Tree class method to the Python module documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12500'>#12500</a>] - Removed references to UDP in the documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12523'>#12523</a>] - Removed Sparkling Water topics from H2O-3 User Guide. These are in the Sparkling Water User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12533'>#12533</a>] - Added a Resources section to the Overview and included links to the awesome-h2o repository, H2O.ai blogs, and customer use cases.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12549'>#12549</a>] - Updated GCP Installation documentation with infomation about quota limits.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12565'>#12565</a>] - Updated Gains/Lift documentation. 16 groups are now used by default.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12611'>#12611</a>] - Added Python examples to the Cross-Validation topic in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12617'>#12617</a>] - Added `loss_by_col` and `loss_by_col_idx` to list of GLRM parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12664'>#12664</a>] - Updated documentation for `class_sampling_factors`. `balance_classes` must be enabled when using `class_sampling_factors`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12691'>#12691</a>] - Added a Python example for initializing and starting h2o-3 in Docker.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12709'>#12709</a>] - Updated the Admin menu documentation in Flow after adding "Download Gen Model" option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12757'>#12757</a>] - In GBM and DRF, `enum_limited` is a supported option for `categorical_encoding`.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12810'>#12810</a>] - Added the -notify_local flag to list of flags available when starting H2O-3 from the command line.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12830'>#12830</a>] - Added documentation for Isolation Forest (beta).
</li>
</ul>
Wright (3.20.0.10) - 10/16/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/10/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/10/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12474'>#12474</a>] - AutoML now correctly. respects the max_runtime_secs setting.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12708'>#12708</a>] - Fixed a multinomial COD solver bug.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12771'>#12771</a>] - Fixed an issue that caused importing of ARFF files to fail if the header was too large and/or with large datasets with categoricals.
</li>
</ul>
Wright (3.20.0.9) - 10/1/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/9/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/9/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12781'>#12781</a>] - Fixed an issue that caused H2O to fail when loading a GLRM model.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12788'>#12788</a>] - log4j.properties can be loaded from classpath.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12789'>#12789</a>] - Buffer configuration is now available for http/https connections.
</li>
</ul>
Wright (3.20.0.8) - 9/21/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/8/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/8/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12707'>#12707</a>] - Fixed an issue that occurred when parsing columns that include double quotation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12732'>#12732</a>] - The `max_runtime_secs` option is no longer ignored when using the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12758'>#12758</a>] - Fixed an XGBoost Sparsity detection test to make it deterministic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12759'>#12759</a>] - Hadoop driver class no longer fails to parse new Java version string.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12713'>#12713</a>] - Added a GBM/DRF Tree walker API in the R client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12714'>#12714</a>] - The R API for obtaining and traversing model trees in GBM/DRF is available in Python.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12562'>#12562</a>] - Added support for user defined split points in partial dependence plots.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12603'>#12603</a>] - Confusion matrices can now be generated in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12752'>#12752</a>] - Java version error messages now reference versions 7 and 8 instead of 1.7 and 1.8.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12754'>#12754</a>] - A Python tree traversal demo is available at <a href="https://github.com/h2oai/h2o-3/blob/master/h2o-py/demos/tree_demo.ipynb">https://github.com/h2oai/h2o-3/blob/master/h2o-py/demos/tree_demo.ipynb</a>.
</li>
</ul>
Wright (3.20.0.7) - 8/31/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/7/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12679'>#12679</a>] - Fixed an issue that caused a mismatch between GLRM MOJO predict and GLRM predict.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12693'>#12693</a>] - Fixed an issue that caused H2O XGBoost grid search to fail even when sizing the sessions 4xs the data size and using extramempercent of 150.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12700'>#12700</a>] - When performing multiple AutoML runs using the H2O R client, viewing the first AutoML leaderboard no longer results in an error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12716'>#12716</a>] - H2O now only binds to the local interface when started from R/Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12723'>#12723</a>] - Fixed an issue that caused DeepLearning and XGBoost MOJOs to get a corrupted input row. This occurred when GenModel's helper functions that perform 1-hot encoding failed to take correctly into considerations cases where useAllFactorLevels = false and corrupted the first categorical value in the input row.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12724'>#12724</a>] - Added gamma, tweedie, and poisson objective functions to the XGBoost Java Predictor.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12729'>#12729</a>] - Fixed an issue in HDFS file import. In rare cases the import could fail due to temporarily inconsistent state of H2O distributed memory.
</li>
</ul>
Wright (3.20.0.6) - 8/24/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/6/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12580'>#12580</a>] - H2oApi.frameColumn in h2o-bindings.jar now correctly parses responses.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12606'>#12606</a>] - biz.k11i:xgboost-predictor:0.3.0 is now ported to the h2oai repo and released to Maven Central. This allows for easier deployment of H2O and Sparkling Water.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12640'>#12640</a>] - In GLM, the coordinate descent solver is now only disabled for when family=multinomial.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12646'>#12646</a>] - Fixed an issue that caused the H2O parser to hang when reading a Parquet file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12657'>#12657</a>] - Fixed an issue that resulted in an AutoML "Unauthorized" Error when running through Enterprise Steam via R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12672'>#12672</a>] - Leaf Node assignment no longer produces the wrong paths for degenerated trees.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12676'>#12676</a>] - Updated the list of Python dependencies on the release download page and in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12679'>#12679</a>] - Fixed an issue that resulted in a mismatch between GLRM predict and GLRM MOJO predict.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12696'>#12696</a>] - Launching H2O on a machine with greater than 2TB no longer results in an integer overflow error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12699'>#12699</a>] - The HTTP parser no longer reads fewer rows when the data is compressed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12703'>#12703</a>] - AstFillNA Rapids expression now returns H2O.unimp() on backward methods.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12590'>#12590</a>] - In GBM and DRF, tree traversal and information is now accessible from the R and Python clients. This can be done using the new h2o.getModelTree function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12634'>#12634</a>] - In GBM, added a new staged_predict_proba function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12666'>#12666</a>] - MOJO output now includes terminal node IDs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12684'>#12684</a>] - GBM/DRF, the H2OTreeClass function now allows you to specify categorical levels.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12697'>#12697</a>] - Updated the XGBoost dependency to ai.h2o:xgboost-predictor:0.3.1.
</li>
</ul>
<h2>Improvement</h2>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12689'>#12689</a>] - Terminal node IDs can now be retrieved in the predict_leaf_node_assignment function.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12688'>#12688</a>] - The User Guide now indicates that only Hive versions 2.2.0 or greater are supported for JDBC drivers. Hive 2.1 is not currently supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12690'>#12690</a>] - In GLM, the documentation for the Coordinate Descent solver now notes that Coordinate Descent is not available when family=multinomial.
</li>
</ul>
Wright (3.20.0.5) - 8/8/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12406'>#12406</a>] - Hive smoke tests no longer time out on HDP.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12647'>#12647</a>] - AutoML now correctly ignores columns specified in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12648'>#12648</a>] - In Flow, the Import SQL Table button now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12660'>#12660</a>] - XGBoost cross validation now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12665'>#12665</a>] - Fixed an issue that caused AutoML to fail in Flow due to the keep_cross_validation_fold_assignment option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12668'>#12668</a>] - Multinomial Stacked Ensemble no longer fails when either XGBoost or Naive Bayes is the base model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12670'>#12670</a>] - Fixed an issue that caused XGBoost to generate the wrong metrics for multinomial cases.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12673'>#12673</a>] - Increased the client_disconnect_timeout value when ClientDisconnectCheckThread searches for connected clients.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12667'>#12667</a>] - Added automated Flow test for AutoML.
</li>
</ul>
Wright (3.20.0.4) - 7/31/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12418'>#12418</a>] - In Flow, increased the height of the summary section for the column summary.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12576'>#12576</a>] - Cross-validation now works correctly in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12594'>#12594</a>] - Documentation for the MOJO predict functions (mojo_predict_pandas and mojo_predict_csv) is now available in the Python User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12599'>#12599</a>] - Regression comparison tests no longer fail between H2OXGBoost and native XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12615'>#12615</a>] - GBM/DRF MOJO scoring no longer allocates unnecessary objects for each scored row.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12591'>#12591</a>] - In GBM, added point estimation as a metric.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12585'>#12585</a>] - Reduced the size of the h2o.jar.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12295'>#12295</a>] - The h2o.importFile([List of Directory Paths]) function will now import all the files located in the specified folders.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12496'>#12496</a>] - Added Standard Error of Mean (SEM) to Partial Dependence Plots.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12574'>#12574</a>] - Added two new formatting options to hex.genmodel.tools.PrintMojo. The --decimalplaces (or -d) option allows you to set the number of places after the decimal point. The --fontsize (or -f) option allows you to set the fontsize. The default fontsize is 14.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12588'>#12588</a>] - Optimized the performance of ingesting large number of small Parquet files by using sequential parse.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12604'>#12604</a>] - Added support for weights in a calibration frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12607'>#12607</a>] - Added a new port_offset command. This parameter lets you specify the relationship of the API port ("web port") and the internal communication port. The previous implementation expected h2o port = api port + 1. Because there are assumptions in the code that the h2o port and API port can be derived from each other, we cannot fully decouple them. Instead, this new option lets the user specify an offset such that h2o port = api port + offset. This enables the user to move the communication port to a specific range, which can be firewalled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12620'>#12620</a>] - Improved speed of ingesting data from HTTP/HTTPS data sources in standalone H2O.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12550'>#12550</a>] - The User Guide now specifies that XLS/XLSX files must be BIFF 8 format. Other formats are not supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12586'>#12586</a>] - Added to docs that when downloading MOJOs/POJOs, users must specify the entire path and not just the relative path.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12629'>#12629</a>] - Added documentation for the new port_offset command when starting H2O.
</ul>
Wright (3.20.0.3) - 7/10/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12221'>#12221</a>] - The `fold_column` option now works correctly in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12423'>#12423</a>] - Calling `describe` on empty H2O frame no longer results in an error in Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15443'>#15443</a>] - In XGBoost, when performing a grid search from Flow, the correct cross validation AUC score is now reported back.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12473'>#12473</a>] - Fixed an issue that cause XGBoost to fail with Tesla V100 drivers 70 and above and with CUDA 9.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12513'>#12513</a>] - H2O's XGBoost results no longer differ from native XGBoost when dmatrix_type="sparse".
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12531'>#12531</a>] - In the R documentation, fixed the description for h2o.sum to state that this function indicates whether to return an H2O frame or one single aggregated sum.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12532'>#12532</a>] - H2O data import for Parquet files no longer fails on numeric decimalTypes.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12542'>#12542</a>] - Fixed an error that occurred when viewing the AutoML Leaderboard in Flow before the first model was completed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12545'>#12545</a>] - When connecting to a Linux H2O Cluster from a Windows machine using Python, the `import_file()` function can now correctly locate the file on the Linux Server.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12548'>#12548</a>] - H2O now reports the project version in the logs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12556'>#12556</a>] - In CoxPH, fixed an issue that caused training to fail to create JSON output when the dataset included too many features.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12563'>#12563</a>] - Users can now switch between edit and command modes on Scala cells.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12577'>#12577</a>] - Fixed an issue with the way that RMSE was calculated for cross-validated models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12583'>#12583</a>] - In GLRM, fixed an issue that caused differences between the result of h2o.predict and MOJO predictions.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12539'>#12539</a>] - Added a new `-report_hostname` flag that can be specified along with `-proxy` when starting H2O on Hadoop. When this flag is enabled, users can replace the IP address with the machine's host name when starting Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12553'>#12553</a>] - Added support for the Amazon Redshift data warehouse.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12581'>#12581</a>] - Added support for CDH 5.9.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12489'>#12489</a>] - Accessing secured (Kerberized) HDFS from a standalone H2O instance works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12515'>#12515</a>] - AutoML Python tests always use max models to avoid running out of time.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12541'>#12541</a>] - CoxPH now validates that a `stop_column` is specified. `stop_column` is a required parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12547'>#12547</a>] - Fixed an issue that caused a GCS Exception to display when H2O was launched offline.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12434'>#12434</a>] - In Flow, improved the display of the confusion matrix for multinomial cases.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12524'>#12524</a>] - Users will now see a Precision-Recall AUC when training binomial models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12525'>#12525</a>] - Synchronous and Asynchronous Scala Cells are now allowed in H2O Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12546'>#12546</a>] - H2O now autodetects string columns and skips them before calculating `groupby`. H2O also warns the user when this happens.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12291'>#12291</a>] - The h2o.mojo_predict_csv and h2o.mojo_predict_df functions now appear in the R HTML documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12558'>#12558</a>] - In GLM, documented that the Poisson family uses the -log(maximum likelihood function) for deviance.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12566'>#12566</a>] - Fixed the R example in the "Replacing Values in a Frame" data munging topic. Columns and rows do not start at 0; R has a 1-based index.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12567'>#12567</a>] - Fixed the R example in the "Group By" data munging topic. Specify the "Month" column instead of the "NumberOfFlights" column when finding the number of flights in a given month based on origin.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12570'>#12570</a>] - Added the new `-report_hostname` flag to the list of Hadoop launch parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12571'>#12571</a>] - Added Amazon Redshift to the list of supported JDBC drivers.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12582'>#12582</a>] - Added CDH 5.9 to the list of supported Hadoop platforms.
</li>
</ul>
Wright (3.20.0.2) - 6/15/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10843'>#10843</a>] - Fixed an issue that resulted in a null pointer exception for H2O ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12122'>#12122</a>] - In AutoML, ignored_columns are now passed in the API call when specifying both x and a fold_column in during training.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12483'>#12483</a>] - Fixed a bug in documentation that incorrectly referenced 'calibrate_frame' instead of 'calibration_frame'.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12490'>#12490</a>] - java -jar h2o.jar no longer fails on Java 7.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12493'>#12493</a>] - Fixed a typo in the AutoML pydocs for sort_metric.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12510'>#12510</a>] - Exported CoxPH functions in R.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12482'>#12482</a>] - Added balance_classes, class_sampling_factors, and max_after_balance_size options to AutoML in Flow.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10651'>#10651</a>] - Updated the project URL, bug reports link, and list of authors in the h2o R package DESCRIPTION file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12405'>#12405</a>] - Update description of the h2o R package in the DESCRIPTION file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12432'>#12432</a>] - AutoML now produces an error message when a response column is missing.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12484'>#12484</a>] - Fixed intermittent test failures for AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12486'>#12486</a>] - Removed frame metadata calculation from AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12494'>#12494</a>] - Removed the keep_cross_validation_models = False argument from the AutoML User Guide examples.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12495'>#12495</a>] - Users can now set a MAX_CM_CLASSES parameter to set a maximum number of confusion matrix classes.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12480'>#12480</a>] - Updated the AutoML screenshot in Flow to show the newly added parameters.
</li>
</ul>
Wright (3.20.0.1) - 6/6/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wright/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wright/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11188'>#11188</a>] - In Scala, the `new H2OFrame()` API no longer fails when using http/https URL-based data sources.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11743'>#11743</a>] - Fixed an issue that caused the Java client JVM to get stuck with a latch/lock leak on the server.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12210'>#12210</a>] - Fixed an issue that caused intermittent NPEs in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12225'>#12225</a>] - In parse, each lock now includes the owner rather than locking with null.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12227'>#12227</a>] - LDAP documentation now contains the correct name of the Auth module.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12293'>#12293</a>] - h2o.jar no longer includes a Jetty 6 dependency.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12328'>#12328</a>] - `model_summary` is now available when running Stacked Ensembles in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12344'>#12344</a>] - XGBoost now correctly respects the H2O `nthreads` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12354'>#12354</a>] - Fixed an invalid invariant in the recall calculation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12358'>#12358</a>] - h2o-genmodel.jar can now be loaded into Spark's spark.executor.extraClassPath.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12367'>#12367</a>] - AutoML now correctly detects the leaderboard frame in H2O Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12388'>#12388</a>] - In XGBoost, fixed an issue that resulted in a "Check failed: param.max_depth < 16 Tree depth too large" error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12414'>#12414</a>] - Zero decimal values and NAs are now represented correctly in XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12415'>#12415</a>] - Response variable datatype checks are now extended to include TIME datatypes.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12459'>#12459</a>] - The `-proxy` argument is now available as part of the h2odriver.args file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12466'>#12466</a>] - Fixed `stopping_metric` values in user guide. Abbreviated values should be specified using upperchase characters (for example, MSE, RMSE, etc.).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12471'>#12471</a>] - Proxy Mode of h2odriver now supports a notification file (specified with the `-notify` argument).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12478'>#12478</a>] - Fixed an issue that caused h2o.predict to throw an exception in H2OCoxPH models with interactions with stratum.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10797'>#10797</a>] - Added MOJO support in Python (via jar file).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11805'>#11805</a>] - Added the `sort_metric` argument to AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11817'>#11817</a>] - Users now have the option to save CV predictions and CV models in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11845'>#11845</a>] - Added an `h2o.H2OFrame.rename` method to rename columns in Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11867'>#11867</a>] - MOJO and POJO support are now available for AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11895'>#11895</a>] - Added support for the Cox Proportional Hazard (CoxPH) algorithm. Note that this is currently available in R and Flow only. It is not yet available in Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12049'>#12049</a>] - Added h2o.get_automl()/h2o.getAutoML function to R/Python APIs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12244'>#12244</a>] - Added the `balance_classes`, `class_sampling_factors`, and max_after_balance_size` arguments to AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12275'>#12275</a>] - When running GLM in Flow, users can now see the InteractionPairs option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12291'>#12291</a>] - Added support for MOJO scoring on a CSV or data frame in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12318'>#12318</a>] - Added an "export model as MOJO" button to Flow for supported algorithms.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12384'>#12384</a>] - Added support for XGBoost MOJO deployment on Windows 10.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12393'>#12393</a>] - GBM and DRF MOJOs and POJOs now return leaf node assignments.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12460'>#12460</a>] - Added the `sort_metric` option to AutoML in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12461'>#12461</a>] - keep_cross_validation_predictions and keep_cross_validation_models are now available when running AutoML in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12476'>#12476</a>] - Deep Learning MOJO now extends Serializable.
</li>
</ul>
<h4>Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12265'>#12265</a>] - In CoxPH, when a categorical column is only used for a numerical-categorical interaction, the algorithm will enforce useAllFactorLevels for that interaction.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11453'>#11453</a>] - When running AutoML and XGBoost, fixed an issue that caused the adapting test frame to be different than the train frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11705'>#11705</a>] - Removed Domain length check for Stacked Ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11931'>#11931</a>] - GLRM predict no longer generates different outputs when performing predictions on training and testing dataframes.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12235'>#12235</a>] - Added support for ingesting data from Hive2 using SQLManager (JDBC interface). Note that this is experimental and is not yet suitable for large datasets.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11264'>#11264</a>] - Replaced the Jama SVD computation in PCA with netlib-java library MTJ.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11402'>#11402</a>] - Created more tests in AutoML to ensure that all fold_assignment values and fold_column work correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11454'>#11454</a>] - Fixed an NPE the occurred when clicking on View button while running AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11463'>#11463</a>] - Bundled Windows XGboost libraries.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11500'>#11500</a>] - Search-based models are no longer duplicated when AutoML is run again on the same dataset with the same seed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11597'>#11597</a>] - When running Stacked Ensembles in R, added support for a vector of base_models in addition to a list.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11833'>#11833</a>] - Added support for Java 9.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12255'>#12255</a>] - Fixed an issue that resulted in an additional progress bar when running h2o.automl() in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12278'>#12278</a>] - Fixed an issue that resulted in an additional progress bar when running AutoML in Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12306'>#12306</a>] - The runint_automl_args.R test now always builds at least 2 models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12325'>#12325</a>] - Improved XGBoost speed by not recreating DMatrix in each iteration (during training).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12342'>#12342</a>] - `offset_column` is now exposed in EasyPredictModelWrapper.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12343'>#12343</a>] - Improved single node XGBoost performance.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12352'>#12352</a>] - Added support for pip 10.0.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12363'>#12363</a>] - In GLM, gamma distribution with 0's in the response results in an improved message: "Response value for gamma distribution must be greater than 0."
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12365'>#12365</a>] - Added metrics to AutoML leaderboard. Binomial models now also show mean_per_class_error, rmse, and mse. Multinomial problems now also show logloss, rmse and mse. Regression models now also show mse.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12397'>#12397</a>] - Exposed `model dump` in XGBoost MOJOs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12401'>#12401</a>] - Improved rebalance for Frames.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12416'>#12416</a>] - Introduced the precise memory allocation algorithm for XGBoost sparse matrices.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12438'>#12438</a>] - Improved SSL documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12462'>#12462</a>] - The Exclude Algorithms section in Flow AutoML is now always visible, even if you have not yet selected a training frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12467'>#12467</a>] - Removes unused parameters, fields, and methods from AutoML. Also exposed buildSpec in the AutoML REST API.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11853'>#11853</a>] - Updated documentation to indicate support for Java 9.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12026'>#12026</a>] - Added the new `pca_impl` parameter to PCA section of the user guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12036'>#12036</a>] - Added a Checkpointing Models section to the User Guide. This describes how checkpointing works for each supported algorithm.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12268'>#12268</a>] - In the "Getting Data into H2O" section, added a link to the new Hive JDBC demo.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12274'>#12274</a>] - The Import File example now also shows how to import from HDFS.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12302'>#12302</a>] - Fixed markdown headings in the example Flows.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12340'>#12340</a>] - All installation examples use H2O version 3.20.0.1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12362'>#12362</a>] - Added a "Data Manipulation" topic for target encoding in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12364'>#12364</a>] - Added new keep_cross_validation_models and keep_cross_validation_predictions options to the AutoML documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12373'>#12373</a>] - Added an example of using XGBoost MOJO with Maven.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12377'>#12377</a>] - In the XGBoost chapter, added information describing how to disable XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12417'>#12417</a>] - When running XGBoost on Hadoop, added a note that users should set -extramempercent to a much higher value.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12440'>#12440</a>] - Added a section for the CoxPH (Cox Proportional Hazards) algorithm.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12442'>#12442</a>] - Added a topic describing how to install H2O-3 from the Google Cloud Platform offering.
</li>
</ul>
Wolpert (3.18.0.11) - 5/24/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/11/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/11/index.html</a>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12445'>#12445</a>] - Enabled Java 10 support for CRAN release.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12446'>#12446</a>] - GLM tests no longer fail on Java 10.
</li>
</ul>
Wolpert (3.18.0.10) - 5/22/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/10/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/10/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12421'>#12421</a>] - Fixed an issue for adding Double.NaN to IntAryVisitor via addValue().
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12422'>#12422</a>] - Removed all code that referenced Google Analytics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12427'>#12427</a>] - Disabled version check in H2O-3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12429'>#12429</a>] - Removed all Google Analytics references and code from Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12430'>#12430</a>] - Removed all Google Analytics references and code from Documentation.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12408'>#12408</a>] - The Security chapter in the User Guide now describes how to enforce system-level command-line arguments in h2odriver when starting H2O.
</li>
</ul>
Wolpert (3.18.0.9) - 5/11/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/9/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/9/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12162'>#12162</a>] - Fixed an issue that caused distributed XGBoost to not be registered in the REST API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12196'>#12196</a>] - Fixed an issue that caused XGBoost to crash due "too many open files."
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12310'>#12310</a>] - Frames are now rebalanced correctly on multinode clusters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12330'>#12330</a>] - Fixed an issue that prevented H2O libraries to load in DBC.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12371'>#12371</a>] - Added more robust checks for Colorama version.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12374'>#12374</a>] - Added more robust checks for Colorama version in H2O Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12382'>#12382</a>] - A response column is no longer required when performing Deep Learning grid search with autoencoder enabled.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12391'>#12391</a>] - Fixed a KeyV3 error message that incorrectly referenced KeyV1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12407'>#12407</a>] - The external backend now stores sparse vector values correctly.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12322'>#12322</a>] - Added a new rank_within_group_by function in R and Python for ranking groups and storing the ranks in a new column.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12366'>#12366</a>] - Improved warning messages in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12400'>#12400</a>] - System administrators can now create a configuration file with implicit arguments of h2odriver and use it to make sure the h2o cluster is started with proper security settings.
</li>
</ul>
Wolpert (3.18.0.8) - 4/19/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/8/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/8/index.html</a>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12331'>#12331</a>] - Release for CRAN submission.
</li>
</ul>
Wolpert (3.18.0.7) - 4/14/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/7/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12351'>#12351</a>] - Fixed a MOJO/POJO scoring issue caused by a serialization bug in EasyPredictModelWrapper.
</li>
</ul>
Wolpert (3.18.0.6) - 4/13/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/6/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12350'>#12350</a>] - In XGBoost, fixed a memory issue that caused training to fail even when running on small datasets.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12307'>#12307</a>] - When files have a Ctr-M character as part of data in the row and Ctr-M also signifies the end of line in that file, it is now parsed correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12324'>#12324</a>] - H2O-3 no longer displays the server version in HTTP response headers.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12326'>#12326</a>] - Updated the Mockito library.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12315'>#12315</a>] - Conda packages are now availabe on S3, enabling installation for users who cannot access anaconda.org.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12339'>#12339</a>] - Added an offset to predictBinomial Easy wrapper.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12099'>#12099</a>] - Updated the AutoML chapter of the User Guide to include a link to H2O World AutoML Tutorials and updated code examples that do not use leaderboard_frame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12323'>#12323</a>] - Fixed links to POJO/MOJO tutorials in the GBM FAQ > Scoring section.
</li>
</ul>
Wolpert (3.18.0.5) - 3/28/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11811'>#11811</a>] - AutoML no longer trains a Stacked Ensemble with only one model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11903'>#11903</a>] - GBM and GLM grids no longer fail in AutoML for multinomial problems.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12138'>#12138</a>] - Users can now merge/sort frames that contain string columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12174'>#12174</a>] - Fixed an issue that occured with multinomial GLM POJO/MOJO models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12203'>#12203</a>] - Users can no longer specify a value of 0 for the col_sample_rate_change_per_level parameter. The value for this parameter must be greater than 0 and <= 2.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12205'>#12205</a>] - The H2O-3 Python client no longer returns an incorrect answer when running a conditional statement.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12233'>#12233</a>] - Added support for CDH 5.14.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12234'>#12234</a>] - Fixed an issue that caused XGBoost to fail when running the airlines dataset on a single-node H2O cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12237'>#12237</a>] - The H2O-3 parser can now handle utf-8 characters that appear in the header.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12261'>#12261</a>] - The H2O-3 parser no longer treats the "Ctr-M" character as an end of line on Linux.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12281'>#12281</a>] - H2O no longer generates a warning when predicting without a weights column.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12269'>#12269</a>] - The AutoML leaderboard no longer prints NaNs for non-US locales.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12107'>#12107</a>] - Added a demo of XGBoost in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12253'>#12253</a>] - Improved the ordinal regression parameter optimization by changing the implementation.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10871'>#10871</a>] - In Flow, improved the vertical scrolling for training and validation metrics for thresholds.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12232'>#12232</a>] - Added more logging regarding the WatchDog client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12250'>#12250</a>] - Replaced unknownCategoricalLevelsSeenPerColumn with ErrorConsumer events in POJO log messages.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12267'>#12267</a>] - Improved the logic that triggers rebalance.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12271'>#12271</a>] - AutoML now uses correct datatypes in the AutoML leaderboard TwoDimTable.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12164'>#12164</a>] - Added ``beta constraints`` and ``prior`` entries to the Parameters Appendix, along with examples in R and Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12236'>#12236</a>] - Added CDH 5.14 to the list of supported Hadoop platforms in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12280'>#12280</a>] - Updated the documenation for the Ordinal ``family`` option in GLM based on the new implementation. Also added new solvers to the documenation: GRADIENT_DESCENT_LH and GRADIENT_DESCENT_SQERR.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12283'>#12283</a>] - Added information about Extremely Randomized Trees (XRT) to the DRF chapter in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12288'>#12288</a>] - On the H2O-3 and Sparkling Water download pages, the link to documentation site now points to the most updated version.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12298'>#12298</a>] - The ``target_encode_create`` and ``target_encode_apply`` are now included in the R HTML documentation.
</li>
</ul>
<h4>Fault</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15497'>#15497</a>] - Fixed an issue that caused SQLManager import to break on cluster with over 100 nodes.
</li>
</ul>
Wolpert (3.18.0.4) - 3/8/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/4/index.html</a>
- Fixed minor release process issue preventing Sparkling Water release.
Wolpert (3.18.0.3) - 3/2/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11975'>#11975</a>] - In Flow, the metalearner_fold_column option now correctly displays a drop-down of column names.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12154'>#12154</a>] - Fixed an issue that caused data import and building models fail when using Flow in IE 11.1944 on Windows 10 Enterprise.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12194'>#12194</a>] - Stacked Ensemble no longer fails when using a grid or list of GLMs as the base models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12071'>#12071</a>] - Fixed an issue that caused an error when during Parquet data ingest.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12204'>#12204</a>] - In Random Forest, added back the distribution and offset_column options for backward compatibility. Note that these options are deprecated and will be ignored if used.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12208'>#12208</a>] - MOJO export to a file now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12211'>#12211</a>] - Fixed an NPE that occurred when checking if a request is Xhr.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11884'>#11884</a>] - Added support for ordinal regression in GLM. This is specified using the `family` option.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12146'>#12146</a>] - Added the exclude_algos option to AutoML in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12179'>#12179</a>] - Added a Leave-One-Out Target Encoding option to the R API. This can help improve supervised learning results when there are categorical predictors with high cardinality. Note that a similar function for Python will be available at a later date.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12195'>#12195</a>] - POJO now logs error messages for all incorrect data types and includes default values rather than NULL when a data type is unexpected.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12212'>#12212</a>] - Moved AutoML to the top of the Model menu in Flow.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12177'>#12177</a>] - In the GLM chapter, added Ordinal to the list of `family` options. Also added Ologit, Oprobit, and Ologlog to the list of `link` options, which can be used with the Ordinal family.
</li>
</ul>
Wolpert (3.18.0.2) - 2/20/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12172'>#12172</a>] - Distributed XGBoost no longer fails silently when expanding a 4G dataset on a 1TB cluster.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12126'>#12126</a>] - Fixed an issue that caused GLM Multinomial to not work properly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12150'>#12150</a>] - In XGBoost, when the first domain of a categorical is parseable as an Int, the remaining columns are not automatically assumed to also be parseable as an Int. As a result of this fix, the default value of categorical_encoding in XGBoost is now AUTO rather than label_encoder.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12166'>#12166</a>] - Fixed an issue that caused XGBoost models to fail to converge when an unknown decimal separator existed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12197'>#12197</a>] - Fixed an issue in ParseTime that led to parse failing.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12184'>#12184</a>] - In the User Guide, the default value for categorical_encoding in XGBoost is now AUTO rather than label_encoder.
</li>
</ul>
Wolpert (3.18.0.1) - 2/12/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11467'>#11467</a>] - Fixed an issue that caused XGBoost binary save/load to fail.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11475'>#11475</a>] - Fixed an issue that caused a Levensthein Distance Normalization Error. Levenstein distance is now implemented directly into H2O.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11985'>#11985</a>] - The Word2Vec Python API for pretrained models no longer requires a training frame. In addition, a new `from_external` option was added, which creates a new H2OWord2vecEstimator based on an external model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12001'>#12001</a>] - Fixed an issue that caused the show function of metrics base to fail to check for a key custom_metric_name and excepts.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12002'>#12002</a>] - The fold column in Kmeans is no longer required to be in x.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12003'>#12003</a>] - The date is now parsed correctly when parsed from H2O-R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12006'>#12006</a>] - In Flow, the scoring history plot is now available for GLM models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12008'>#12008</a>] - The Parquet parser no longer fails if one of the files to parse has no records.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12017'>#12017</a>] - Added error checking and logging on all the uses of `water.util.JSONUtils.parse().
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12027'>#12027</a>] - In AutoML, fixed an exception in Python binding that occurred when the leaderboard was empty.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12028'>#12028</a>] - In AutoML, fixed an exception in R binding that occurred when the leaderboard was empty.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12031'>#12031</a>] - Removed Pandas dependency for AutoML in Python.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12039'>#12039</a>] - In PySparkling, reading Parquet/Orc data with time type now works correctly in H2O.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12046'>#12046</a>] - Fixed a maximum recursion depth error when using `isin` in the H2O Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12047'>#12047</a>] - When running getJobs in Flow, fixed a ClassNotFoundException that occurred when AutoML jobs existed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12051'>#12051</a>] - Fixed an issue that caused a list of columns to be truncated in PySparkling. Light endpoint now returns all columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12058'>#12058</a>] - In AutoML, fixed a deadlock issue that occurred when two AutoML runs came in the same second, resulting in matching timestamps.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12063'>#12063</a>] - The offset_column and distribution parameters are no longer available in Random Forest.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12067'>#12067</a>] - Fixed an issue in XGBoost that caused MOJOs to fail to work without manually adding the Commons Logging dependency.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12075'>#12075</a>] - Fixed an issue that caused XGBoost to mangle the domain levels for datasets that have string response domains.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12085'>#12085</a>] - In Flow, the separator drop down now shows 3-digit decimal values instead of 2.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12087'>#12087</a>] - Users can now specify interactions when running GLM in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12100'>#12100</a>] - FrameMetadate code no longer uses hardcoded keys. Also fixed an issue that caused AutoML to fail when multiple AutoMLs are run simultaneously.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12101'>#12101</a>] - A frame can potentially have a null key. If there is a Frame with a null key (just a container for vecs), H2O no longer attempts to track a null key.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12128'>#12128</a>] - Users can now successfully build an XGBoost model as compile chain. XGBoost no longer fails to provide the compatible artifact for an Oracle Linux environment.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12137'>#12137</a>] - GLM no longer fails when a categorical column exists in the dataset along with an empty value on at least one row.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12158'>#12158</a>] - Fixed an issue that cause GBM grid to fail on some datasets when specifying `sample_rate` in the grid.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12159'>#12159</a>] - The x argument is no longer required when performing a grid search.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12168'>#12168</a>] - Fixed an issue that caused the Parquet parser to fail on Spark 2.0 (SW-707).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12186'>#12186</a>] - Fixed an issue that caused XGBoost OpenMP to fail on Ubuntu 14.04.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11000'>#11000</a>] - Added support for INT96 timestamp to the Parquet parser.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11535'>#11535</a>] - Added support for XGBoost multinode training in H2O. Note that this is still a BETA feature.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11856'>#11856</a>] - Users can now specify a list of algorithms to exclude during an AutoML run. This is done using the new `exclude_algos` parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12076'>#12076</a>] - In GLM, users can now specify a list of interactions terms to include when building a model instead of relying on the default action of including all interactions.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12102'>#12102</a>] - The Python PCA code examples in github and in the User Guide now use the h2o.estimators.pca.H2OPrincipalComponentAnalysisEstimator method instead of the h2o.transforms.decomposition.H2OPCA method.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12123'>#12123</a>] - Upgraded the XGBoost version. This now supports RHEL 6.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11959'>#11959</a>] - Stacked Ensemble allows you to specify the metalearning algorithm to use when training the ensemble. When an algorithm is specified, Stacked Ensemble runs with the specified algorithm's default hyperparameter values. The new ``metalearner_params`` option allows you to pass in a dictionary/list of hyperparameters to use for that algorithm instead of the defaults.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12096'>#12096</a>] - Users can now specify a seed parameter in Stacked Ensemble.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12181'>#12181</a>] - Documented clouding behavior of an H2O cluster. This is available at https://github.com/h2oai/h2o-3/blob/master/h2o-docs/devel/h2o_clouding.rst.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12021'>#12021</a>] - Updated the documentation to indicate that datetime parsing from R and Flow now is UTC by default.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12023'>#12023</a>] - R documentation on docs.h2o.ai is now available in HTML format.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12044'>#12044</a>] - Added a new Cloud Integration topic for using H2O with AWS.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12093'>#12093</a>] - In the XGBoost chapter, added that XGBoost in H2O supports multicore.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12114'>#12114</a>] - Added `interaction_pairs` to the list of GLM parameters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12155'>#12155</a>] - Added `metalearner_algorithm` and `metalearner_params` to the Stacked Ensembles chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12182'>#12182</a>] - The H2O-3 download site now includes a link to the HTML version of the R documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12183'>#12183</a>] - Updated the XGBoost documentation to indicate that multinode support is now available as a Beta feature.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12185'>#12185</a>] - Added the seed parameter to the Stacked Ensembles section of the User Guide.
</li>
</ul>
Wheeler (3.16.0.4) - 1/15/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/4/index.html</a>
<h4>Bug</h4>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12078'>#12078</a>] - Fixed several client deadlock issues.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12084'>#12084</a>] - When verifying that a supported version of Java is available, H2O no longer checks for version 1.6.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12088'>#12088</a>] - The H2O-3 download site has an updated link for the Sparkling Water README.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12092'>#12092</a>] - In Aggregator, fixed the way that a created mapping frame is populated.
</li>
<h4>New Feature</h4>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12081'>#12081</a>] - XGBoost can now be used in H2O on Hadoop with a single node.
</li>
<h4>Improvement</h4>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12082'>#12082</a>] - Deep Water is disabled in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12083'>#12083</a>] - This release of H2O includes an upgraded XGBoost version.
</li>
Wheeler (3.16.0.3) - 1/8/2018
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/3/index.html</a>
<h4>Technical task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12056'>#12056</a>] - H2O-3 now allows definition of custom function directly in Python notebooks and enables iterative updates on defined functions.
</li>
</ul>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11740'>#11740</a>] - When a frame name includes numbers followed by alphabetic characters (for example, "250ML"), Rapids no longer parses the frame name as two tokens.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11775'>#11775</a>] - Fixed an issue that caused Partial Dependence Plots to a use different order of categorical values after calling as.factor.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12020'>#12020</a>] - Added support for CDH 5.13.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12052'>#12052</a>] - Fixed an issue that caused a Python 2 timestamp to be interpreted as two tokens.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12068'>#12068</a>] - Aggregator supports categorial features. Fixed a discrepency in the Aggregator documentation.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11504'>#11504</a>] - In GBM, users can now specify quasibinomial distribution.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11842'>#11842</a>] - H2O-3 now supports the Netezza JDBC driver.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12043'>#12043</a>] - Users can now optionally export the mapping of rows in an aggregated frame to that of the original raw data.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11993'>#11993</a>] - When using S3/S3N, revised the documentation to recommend that S3 should be used for data ingestion, and S3N should be used for data export.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12022'>#12022</a>] - The H2O User Guide has been updated to indicate support for CDH 5.13.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12034'>#12034</a>] - Updated the Anaconda section with information specifically for Python 3.6 users.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12050'>#12050</a>] - The H2O User Guide has been updated to indicate support for the Netezza JDBC driver.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12062'>#12062</a>] - Added "quasibinomial" to the list of `distribution` options in GBM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12064'>#12064</a>] - Added the new `save_mapping_frame` option to the Aggregator documentation.
</li>
</ul>
Wheeler (3.16.0.2) - 11/30/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/2/index.html</a>
<h4>Bug</h4>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11988'>#11988</a>] - In AutoML, fixed an issue that caused the leaderboard_frame to be ignored when nfolds > 1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11990'>#11990</a>] - Improved the warning that displays when mismatched jars exist.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11999'>#11999</a>] - The correct H2O version now displays in setup.py for sdist.
</li>
<h4>Improvement</h4>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11984'>#11984</a>] - Incorporated final improvements to the Sparkling Water booklet.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12000'>#12000</a>] - Automated Anaconda releases.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12004'>#12004</a>] - This version of H2O introduces light rest endpoints for obtaining frames in the python client.
</li>
Wheeler (3.16.0.1) - 11/24/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-wheeler/1/index.html</a>
<h4>Technical Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11960'>#11960</a>] - A backend Java API is now available for custom evaluation metrics.
</li>
</ul>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14436'>#14436</a>] - Users can now save models to and download models from S3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10476'>#10476</a>] - When running h2o.merge in the R client, the status line indicator will no longer return quickly. Users can no longer enter new commands until the merge process is completed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11061'>#11061</a>] - In the R client strings, training_frame says no longer states that it is an optional parameter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11553'>#11553</a>] - The H2OFrame.mean method now works in Python 3.6.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11576'>#11576</a>] - Early stopping now works with perfectly predictive data.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11606'>#11606</a>] - h2o.group_by now works correctly when specifying a median() value.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11657'>#11657</a>] - In XGBoost fixed an issue that caused prediction on a dataset without a response column to return an error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11731'>#11731</a>] - When running AutoML in Flow, users can now specify a project name.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11735'>#11735</a>] - h2odriver in proxy mode now correctly forwards the authentication headers to the H2O node.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11778'>#11778</a>] - H2O can ingest Parquet 1.8 files created by Spark.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11784'>#11784</a>] - Loading models and exporting models to/from AWS S3 now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11785'>#11785</a>] - Fixed an issue that caused binary model imports and exports from/to S3 to fail.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11808'>#11808</a>] - Users can now load data from s3n resources after setting core-site.xml correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11831'>#11831</a>] - Fixed an error that occurred when exporting data to s3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11861'>#11861</a>] - Fixed an issue that caused H2O to "forget" that a column is of factor type if it contains only NA values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11872'>#11872</a>] - The download instructions for Python now indicate that version 3.6 is supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11878'>#11878</a>] - In Flow, fixed an issue with retaining logs from the client node.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11879'>#11879</a>] - H2O can now handle the case where I'm the Client and the md5 should be ignored.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11881'>#11881</a>] - h2o.residual_deviance now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11893'>#11893</a>] - h2o.predict no longer returns an error when the user does not specify an offset_column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11908'>#11908</a>] - Fixed an issue with Spark string chunks.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11912'>#11912</a>] - Logs now display correctly on HADOOP, and downloaded logs no longer give an empty folder when the cluster is up.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11913'>#11913</a>] - Added an option for handling empty strings. If compare_empty if set to FALSE, empty strings will be handled as NaNs.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11915'>#11915</a>] - HTTP logs can now be obtained in Flow UI.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11742'>#11742</a>] - Fixed an issue with the progress bar that occurred when running PySparkling + DataBricks.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11940'>#11940</a>] - Fixed reporting of clients with the wrong md5.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11943'>#11943</a>] - In the R and Python clients, updated the strings for max_active_predictors to indicate that the default is now 5000.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11945'>#11945</a>] - h2o.merge now works correctly for one-to-many when all.x=TRUE.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11947'>#11947</a>] - Fixed an issue that caused GLM predict to fail when a weights column was not specified.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11954'>#11954</a>] - Reduced the number of URLs that get sent to google analytics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11968'>#11968</a>] - When building a Stacked Ensemble model, the fold_column from AutoML is now piped through to the stacked ensemble.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11969'>#11969</a>] - Fixed an issue that cause GLM scoring to produce incorrect results for sparse data.
</li>
</ul>
<h4>Epic</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11563'>#11563</a>] - This version of H2O includes support for Python 3.6.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10773'>#10773</a>] - MOJOs are now supported for Stacked Ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10640'>#10640</a>] - User can now specify the metalearner algorithm type that StackedEnsemble should use. This can be AUTO, GLM, GBM, DRF, or Deep Learning.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10864'>#10864</a>] - Added a metalearner_folds option in Stacked Ensembles, enabling cross validation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10975'>#10975</a>] - In GBM, endpoints are now exposed that allow for custom evaluation metrics.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11760'>#11760</a>] - When running AutoML through the Python or R clients, users can now specify the nfolds argument.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11770'>#11770</a>] - Add another Stacked Ensemble (top model for each algo) to AutoML
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11944'>#11944</a>] - The AutoML leaderboard now uses cross-validation metrics (new default).
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11792'>#11792</a>] - K-Means POJOs and MOJOs now expose distances to cluster centers.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11834'>#11834</a>] - Multiclass stacking is now supported in AutoML. Removed the check that caused AutoML to skip stacking for multiclass.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11918'>#11918</a>] - Users can now specify a number of folds when running AutoML in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11957'>#11957</a>] - Added a metalearner_fold_column option in Stacked Ensembles, allowing for custom folds during cross validation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11870'>#11870</a>] - The Aggregator Function is now exposed in the R client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11871'>#11871</a>] - The Aggregator Function is now available in the Python client.
</li>
</ul>
<h4>Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11919'>#11919</a>] - Fixed a Jaro-Winkler Dependency.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11682'>#11682</a>] - The current version of h2o-py is now published into PyPi.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11774'>#11774</a>] - Change behavior of auto-generation of validation and leaderboard frames in AutoML
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11809'>#11809</a>] - Updated the download site and the end user documentation to indicate that Python3.6 is now supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11813'>#11813</a>] - PyPi/Anaconda descriptors now indicate support for Python 3.6.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11670'>#11670</a>] - Enabled the lambda search for the GLM metalearner in Stacked Ensembles. This is set to TRUE and early_stopping is set to FALSE.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11710'>#11710</a>] - Running `pip install` now installs the latest version of H2O-3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11840'>#11840</a>] - In EasyPredictModelWrapper, preamble(), predict(), and fillRawData() are now protected rather than private.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11955'>#11955</a>] - MOJOs/POJOs will not be created for unsupported categorical_encoding values.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11982'>#11982</a>] - An AutoML run now outputs two StackedEnsemble model IDs. These are labeled StackedEnsemble_AllModels and StackedEnsemble_BestOfFamily.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11187'>#11187</a>] - In the Data Manipulation chapter, added a topic for pivoting tables.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11544'>#11544</a>] - Added a topic to the Data Manipulation chapter describing the h2o.fillna function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11626'>#11626</a>] - Added MOJO and POJO Quick Start sections directly into the Productionizing H2O chapter. Previously, this chapter included links to quick start files.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11689'>#11689</a>] - In the GBM booklet when describing nbins_cat, clarified that factors rather than columns get grouped together.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11695'>#11695</a>] - The description for the GLM lambda_max option now states that this is the smallest lambda that drives all coefficients to zero.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11712'>#11712</a>] - Updated the installation instructions for PySparkling.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11741'>#11741</a>] - Clarified that in H2O-3, sampling is without replacement.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11756'>#11756</a>] - Updated documentation to state that multiclass classification is now supported in Stacked Ensembles.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11757'>#11757</a>] - Updated documentation to state that multiclass stacking is now supported in AutoML.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11773'>#11773</a>] - Added an Early Stopping section the Algorithms > Common chapter.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11823'>#11823</a>] - Added a note in Word2vec stating that binary format is not supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11824'>#11824</a>] - In the Parameters Appendix, updated the description for histogram_type=random.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11835'>#11835</a>] - In the Using Flow > Models > Run AutoML section, updated the AutoML screenshot to show the new Project Name field.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11848'>#11848</a>] - Added a Sorting Columns data munging topic describing how to sort a data frame by column or columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11876'>#11876</a>] - In KMeans, updated the list of model summary statistics and training metrics that are outputted.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11887'>#11887</a>] - Removed SortByResponse from the list of categorical_encoding options for Aggregator and K-Means.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11901'>#11901</a>] - Updated the Sparkling Water links on docs.h2o.ai to point to the latest release.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11907'>#11907</a>] - Added a section in the Algorithms chapter for Aggregator.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11929'>#11929</a>] - Updated the description for Save and Loading Models to indicate that H2O binary models are not compatible across H2O versions.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11930'>#11930</a>] - Added ignored_columns and 'x' parameters to AutoML section. Also added the 'x' parameter to the Parameters Appendix.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11935'>#11935</a>] - In DRF, add FAQs describing splitting criteria.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11958'>#11958</a>] - Added the new metalearner_folds and metalearner_fold_assignment parameters to the Defining a Stacked Ensemble Model section in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11962'>#11962</a>] - Updated the Sparking Water booklet. (Also 11962.)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11965'>#11965</a>] - Added the new metalearner_algorithm parameter to Defining a Stacked Ensemble Model section in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11970'>#11970</a>] - The User Guide and the POJO/MOJO Javadoc have been updated to indicate that MOJOs are supported for Stacked Ensembles.
</li>
</ul>
Weierstrass (3.14.0.7) - 10/20/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/7/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11863'>#11863</a>] - h2o.H2OFrame.any() and h2o.H2OFrame.all() not working properly if frame contains only True
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11864'>#11864</a>] - Don&39;t check H2O client hash-code ( Fix )
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10896'>#10896</a>] - Generate Python API tests for Python Module Data in H2O and Data Manipulation
</li>
</ul>
Weierstrass (3.14.0.6) - 10/9/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/6/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/sparkling-water/issues/4170'>sparkling-water-#4170</a>] - Fixed an issue that prevented Sparkling Water from importing Parquet files.
</li>
</ul>
Weierstrass (3.14.0.5) - 10/9/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/5/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11748'>#11748</a>] - Fixed an issue that caused sorting to be done incorrectly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11795'>#11795</a>] - Only relevant clients (the ones with the same cloud name) are now reported to H2O. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11832'>#11832</a>] - Improved error messaging in the case where H2O fails to parse a valid Parquet file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11836'>#11836</a>] - Fixed an issue that allowed nodes from different clusters to kill different H2O clusters.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11855'>#11855</a>] - Fixed an issue that caused K-Means to improperly calculate scaled distance.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11803'>#11803</a>] - Nightly and stable releases will now have published sha256 hashes.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11290'>#11290</a>] - The h2o.sort() function now includes an `ascending` parameter that allows you to specify whether a numeric column should be sorted in ascending or descending order.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11841'>#11841</a>] - H2O no longer terminates when an incompatible client tries to connect.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11827'>#11827</a>] - Updated the list of required packages for the H2O-3 R client on the H2O Download site and in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11843'>#11843</a>] - Added an FAQ to the <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/faq/java.html">User Guide FAQ</a> describing how Java 9 users can switch to a supported Java version.
</li>
</ul>
Weierstrass (3.14.0.3) - 9/18/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/3/index.html</a>
<h4>Technical Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11751'>#11751</a>] - Introduced a Python client side AST optimization.
</li>
</ul>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10437'>#10437</a>] - In R, `h2o.arrange()` can now sort on a float column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11602'>#11602</a>] - The `as_data_frame()` function no longer drops rows with NAs when `use_pandas` is set to TRUE.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11614'>#11614</a>] - In Deep Learning POJOs, fixed an issue in the sharing stage between threads.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11618'>#11618</a>] - Fixed an issue in R that caused `h2o.sub` to fail to retain the column names of the frame.
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11636'>#11636</a>] - Running ifelse() on a constant column no longer results in an error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11725'>#11725</a>] - Using + on string columns now works correctly.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11727'>#11727</a>] - Fixed an issue that caused a POJO and a MOJO to return different column names with the `getNames()` method.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11728'>#11728</a>] - The R and Python clients now have consistent timeout numbers.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11746'>#11746</a>] - Fixed an issue that resulted in an AIOOB error when predicting with GLM. NA responses are now removed prior to GLM scoring.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11787'>#11787</a>] - The set_name method now works correctly in the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11799'>#11799</a>] - Replaced the deprecated Clock class in timing.gradle.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11815'>#11815</a>] - The MOJO Reader now closes open files after reading.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11510'>#11510</a>] - MOJO support has been extended to include the Deep Learning algorithm.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11724'>#11724</a>] - Added the ability to import an encrypted (AES128) file into H2O. This can be configured glovally by specifying the `-decrypt_tool` option and installing the tool in DKV.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11782'>#11782</a>] - The Decryption API is now exposed in the REST API and in the R client.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11690'>#11690</a>] - Updated the MOJO Quick Start Guide to show separator differences between Linux/OS X and Windows. Also updated the R example to match the Python example.
</li>
</ul>
Weierstrass (3.14.0.2) - 8/21/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11683'>#11683</a>] - Fixed a broken link to the Hive tutorials from the Productionizing section in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11701'>#11701</a>] - Sparkling Water can now pass a data frame with a vector for conversion into H2OFrame. In prior versions, the vector was not properly expanded and resulted in a failure.
</li>
</ul>
<h4>Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11681'>#11681</a>] - Added more tests to ensure that, when max_runtime_secs is set, the returned model works correctly.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11691'>#11691</a>] - This version of H2O includes an option to force toggle (on/off) a specific extension. This enables users to enable the XGBoost REST API on a system that does not support XGBoost.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11708'>#11708</a>] - A warning now displays when the minimal XGBoost version is used.
</li>
</ul>
Weierstrass (3.14.0.1) - 8/10/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/1/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9702'>#9702</a>] - In the R client, making a copy of a factor column and then changing the factor levels no longer causes the levels of the original column to change.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11466'>#11466</a>] - Added a **Leaderboard Frame** option in Flow when configuring an AutoML run.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11468'>#11468</a>] - The `h2o.performance` function now works correctly on XGBoost models.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11507'>#11507</a>] - In the Python client, improved the help string for `h2o_import_file`. This string now indicates that setting `(parse=False)` will return a list instead of an H2OFrame.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11537'>#11537</a>] - Removed the Ecko dependency. This is not needed.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11562'>#11562</a>] - Fixed an issue that caused the parquet parser to store numeric/float values in a string column. This issue occurred when specifying an unsupported type conversion in Parse Setup (for example, numeric -> string). Users will now encounter an error when attempting this. Additionally, users can now change Enums->Strings in parse setup.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11565'>#11565</a>] - Deep Learning POJOs are now thread safe.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11567'>#11567</a>] - Fixed the default print method for H2OFrame in Python. Now when a user types the H2OFrame name, a new line is added, and the header is pushed to the next line.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11581'>#11581</a>] - Fixed an issue that caused the `max_runtime_secs` parameter to fail correctly when run through the Python client. As a result of this fix, the `max_runtime_secs` parameter was added to Word2vec.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11583'>#11583</a>] - Fixed an issue that caused XGBoost grid search to fail when using the Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11603'>#11603</a>] - When running with weighted data and columns that are constant after applying weights, a GLM lambda search no longer results in an AIOOB error.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11609'>#11609</a>] - The XGBoost `max_bin` parameter has been renamed to `max_bins`, and its default value is now 256.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11610'>#11610</a>] - XGBoost Python documentation is now available.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11611'>#11611</a>] - In XGBoost, the `learning_rate` (alias: `eta` parameter now has a default value of 0.3.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11613'>#11613</a>] - In XGBoost, the `max_depth` parameter now has a default value of 6.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11614'>#11614</a>] - Multi-threading is now supported by POJO downloaded.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11630'>#11630</a>] - The XGBoost `min_rows` (alias: `min_child_weight`) parameter now has a default value of 1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11631'>#11631</a>] - The XGBoost `max_abs_leafnode_pred` (alias: `max_delta_step`) parameter now has a default value of 0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11632'>#11632</a>] - H2O XGBoost default options are now consistent with XGBoost default values. This fix involved the following changes:
<ul>
<li>num_leaves has been renamed max_leaves, and its default value is 0.
</li>
<li>The default value for reg_lambda is 0.
</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11635'>#11635</a>] - Removed the Guava dependency from the Deep Water API.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11655'>#11655</a>] - In XGBoost, the default value for sample_rate and the alias subsample are now both 1.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11656'>#11656</a>] - In XGBoost, the default value for colsample_bylevel (alias: colsample_bytree) has been changed to 1.0.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11662'>#11662</a>] - Hidden files are now ignored when reading from HDFS.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11331'>#11331</a>] - Added a `verbose` option to Deep Learning, DRF, GBM, and XGBoost. When enabled, this option will display scoring histories as a model job is running.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11561'>#11561</a>] - Added an `extra_classpath` option, which allows users to specify a custom classpath when starting H2O from the R and Python client.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11564'>#11564</a>] - Users can now override the type of a Str/Cat column in a Parquet file when the parser attempts to auto detect the column type.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11617'>#11617</a>] - Users can now run a standalone H2O instance and read from a Kerberized cluster's HDFS.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11624'>#11624</a>] - Added support for CDH 5.10.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11629'>#11629</a>] - Added support for MapR 5.2.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10840'>#10840</a>] - Fixed an issue that caused PCA to take 39 minutes to run on a wide dataset.
The wide dataset method for PCA is now only enabled if the dataset is very wide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11478'>#11478</a>] - XGBoost-specific WARN messages have been converted to TRACE.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11506'>#11506</a>] - When printing frames via `head()` or `tail()`, the `nrows` option now allows you to specify more than 10 rows. With this change, you can print the complete frame, if desired.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11512'>#11512</a>] - Improved the speed of converting a sparse matrix to an H2OFrame in R.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11546'>#11546</a>] - Added the following parameters to the XGBoost R/Py clients:
<ul>
<li>categorical_encoding</li>
<li>sample_type</li>
<li>normalize_type</li>
<li>rate_drop</li>
<li>one_drop</li>
<li>skip_drop</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11557'>#11557</a>] - H2O can now handle sparse vectors as the input of the external frame handler.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11571'>#11571</a>] - Added MOJO support for Spark SVM.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11580'>#11580</a>] - When running AutoML from within Flow, the default `stopping_tolerance` is now NULL instead of 0.001.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11627'>#11627</a>] - Removed dependency on Reflections.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11406'>#11406</a>] - Updated the list of Python requirements in the <a href="https://github.com/h2oai/h2o-3/blob/master/README.md#42-setup-on-all-platforms">README.md</a>, on the download site, and in the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11437'>#11437</a>] - Updated the FAQ for Saving and Loading a Model in K-Means.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11449'>#11449</a>] - Added a <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/flow.html#run-automl">Run AutoML</a> subsection in the Flow section of the User Guide.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11482'>#11482</a>] - Continued improvements to XGBoost documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11511'>#11511</a>] - Added documentation for using H2O SW with Databricks.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11514'>#11514</a>] - In the <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/faq/general.html">http://docs.h2o.ai/h2o/latest-stable/h2o-docs/faq/general.html</a> topic, updated the example for scoring using an exported POJO.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11531'>#11531</a>] - In the <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/productionizing.html#about-pojos-and-mojos">About POJOs and MOJOs</a> topic, added text describing the h2o-genmodel jar file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11539'>#11539</a>] - The User Guide now indicates that Hive files can be saved in ORC format and
then imported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11568'>#11568</a>] - For topics that indicate support for Avro-formatted data, updated the User Guide to reflect that only Avro version 1.8.0 is supported.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11599'>#11599</a>] - A new H2O Python / Pandas Munging Parity document is now available at <a href="https://github.com/h2oai/h2o-3/tree/master/h2o-docs/src/cheatsheets">https://github.com/h2oai/h2o-3/tree/master/h2o-docs/src/cheatsheets</a>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11612'>#11612</a>] - Added parameter defaults to the <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/xgboost.html">XGBoost</a> section in the User Guide.
</li>
</ul>
Vapnik (3.12.0.1) 6/6/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-vapnik/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-vapnik/1/index.html</a>
<h4>Epic</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11162'>#11162</a>] - AutoML is now available in H2O. AutoML can be used for automatically training and tuning a number of models within a user-specified time limit or model limit. It is designed to run with as few parameters as possible, and the top performing models can be viewed on a leaderboard. More information about AutoML is available <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/automl.html'>here</a>.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11336'>#11336</a>] - With the addition of the AutoML feature, a new **Run AutoML** option is available in Flow under the **Models** dropdown menu.</li>
</ul>
Vajda (3.10.5.4) - 7/17/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/4/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11573'>#11573</a>] - Fixed an issue that caused tree algos to waste memory by storing categorical values in every tree.
</li>
</ul>
Vajda (3.10.5.3) - 6/30/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/3/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10917'>#10917</a>] - Fixed an issue that resulted in "Unexpected character after column id:" warnings when parsing an SVMLight file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11330'>#11330</a>] - h2o.predict now displays a warning if the features (columns) in the test frame do not contain those features used by the model.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11455'>#11455</a>] - The XGBoost REST API is now only registered when backend lib exists.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11477'>#11477</a>] - H2O no longer displays an error if there is a "/" in the user-supplied model name. Instead, a message will display indicating that the "/" is replaced with "_".
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10835'>#10835</a>] - Added support for autoencoder POJOs in in the EasyPredictModelWrapper.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11158'>#11158</a>] - H2O now warns the user about the minimal required Colorama version in case of python client. Note that the current minimum version is 0.3.8.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11421'>#11421</a>] - Removed deprecation warnings from the H2O build.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11432'>#11432</a>] - Moved the initialization of XGBoost into the H2O core extension.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11399'>#11399</a>] - Added a link to paper describing balance classes in the <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/algo-params/balance_classes.html'>balance_classes</a> parameter topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11492'>#11492</a>] - Removed `laplace`, `huber`, and `quantile` from list of supported distributions in the XGBoost documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11494'>#11494</a>] - Add heuristics to the FAQ > General Troubleshooting topic.
</li>
</ul>
Vajda (3.10.5.2) - 6/19/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/2/index.html</a>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10756'>#10756</a>] - In PCA, fixed an issue that resulted in errors when specifying `pca_method=glrm` on wide datasets. In addition, the GLRM algorithm can now be used with wide datasets.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11302'>#11302</a>] - Fixed issues with streamParse in ORC parser that caused a NullPointerException when parsing multifile from Hive.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11323'>#11323</a>] - Fixed an issue that occurred with H2O data frame indexing for large indices that resulted in off-by-one errors. Now, when indexing is set to a value greater than 1000, indexing between left and right sides is no longer inconsistent.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11341'>#11341</a>] - In DRF, fixed an issue that resulted in an AssertionError when run on certain datasets with weights.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9751'>#9751</a>] - Removed an incorrect Python example from the Sparkling Water booklet. Python users must start Spark using the H2O pysparkling egg on the Python path. Using `--package` when running the pysparkling app is not advised, as the pysparkling distribution already contains the required jar file.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11476'>#11476</a>] - In GLM fixed an issue that caused a Runtime exception when specifying the quasibinomial family with `nfold = 2`.
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10529'>#10529</a>] - Added top an bottom N functions, which allow users to grab the top or bottom N percent of a numerical column. The returned frame contains the original row indices of the top/bottom N percent values extracted into the second column.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10985'>#10985</a>] - When building Stacked Ensembles in R, the base_models parameter can accept models rather than just model IDs. Updated the documentation in the User Guide for the base_models parameter to indicate this.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11407'>#11407</a>] - Added the following new GBM and DRF parameters to the User Guide: `calibrate_frame` and `calibrate_model`.
</li>
</ul>
<h4>Improvement</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11415'>#11415</a>] - Improved PredictCsv.java as follows:
<ul>
<li>Enabled PredictCsv.java to accept arbitrary separator characters in the input dataset file if the user includes the optional flag `--separator` in the input arguments. If a user enters a special Java character as the separator, then H2O will add "\".
</li>
<li>Enabled PredictCsv.java to perform setConvertInvalidNumbersToNa(setInvNumNA)) if the optional flag `--setConvertInvalidNum` is included in the input arguments.
</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11461'>#11461</a>] - Fixed the R package so that a "browseURL" NOTE no longer appears.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11465'>#11465</a>] - In the R package documentation, improved the description of the GLM `alpha` parameter.
</li>
</ul>
<h4>Docs</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11408'>#11408</a>] - In the "Using Flow - H2O’s Web UI" section of the User Guide, updated the Viewing Models topic to include that users can download the h2o-genmodel.jar file when viewing models in Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11433'>#11433</a>] - The `group_by` function accepts a number of aggregate options, which were documented in the User Guide and in the Python package documentation. These aggregate options are now described in the R package documentation.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11458'>#11458</a>] - Added an initial XGBoost topic to the User Guide. Note that this is still a work in progress.
</li>
</ul>
Vajda (3.10.5.1) - 6/9/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-vajda/1/index.html</a>
<h4>Technical Task</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14554'>#14554</a>] - Fixed a GLM persist test.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11014'>#11014</a>] - Disabled R tests for OS X.
</li>
</ul>
<h4>Bug</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14428'>#14428</a>] - PCA no longer reports incorrect values when multiple eigenvectors exist.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14542'>#14542</a>] - Users can now specify the weights_column as a numeric index in R.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14549'>#14549</a>] - Fixed an issue that caused GLM models returned by h2o.glm() and h2o.getModel(..) to be different.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14586'>#14586</a>] - Fixed an issue that caused PCA with GLRM to display incorrect results on data.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15193'>#15193</a>] - Fixed an issue that caused `df.show(any_int)` to always display 10 rows.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15321'>#15321</a>] - Starting an H2O cloud from R no longer results in "Error in as.numeric(x["max_mem"]) : (list) object cannot be coerced to type &39;double&39;"
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9595'>#9595</a>] - `h2o::ifelse` now handles NA values the same way that `base::ifelse` does.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9653'>#9653</a>] - Fixed an issue in PCA that resulted in incorrect standard deviation and components results for non standardized data.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9696'>#9696</a>] - When performing a grid search with a `fold_assignment` specified and with `cross_validation` disabled, Python unit tests now display a Java error message. This is because a fold assignment is meaningless without cross validation.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9750'>#9750</a>] - The Python `h2o.get_grid()` function is now in the base h2o object, allowing you to use it the same way as `h2o.get_model()`, `h2o.get_frame()` etc.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10116'>#10116</a>] - The `.mean()` function can now be applied to a row in `H2OFrame.apply()`.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10262'>#10262</a>] - Fixed an issue that caused a negative value to display in the H2O cluster version. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10306'>#10306</a>] - GLM now checks to see if a response is encoded as a factor and warns the user if it is not. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10381'>#10381</a>] - Fixed an issue that resulted in an `h2o.init()` fail message even though the server had actually been started. As a result, H2O did not shutdown automatically upon exit.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10413'>#10413</a>] - Fixed an issue that caused PCA to hang when run on a wide dataset using the Randomized `pca_method`. Note that it is still not recommended to use Randomized with wide datasets.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10432'>#10432</a>] - `h2o.setLevels` now works correctly when wrapped into invisible.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10555'>#10555</a>] - Added a dependency for the roxygen2 package.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10611'>#10611</a>] - `h2o.coef` in R is now functional for multinomial models.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10628'>#10628</a>] - When converting a column to `type = string` with `.ascharacter()` in Python, the `structure` method now correctly recognizes the change.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10657'>#10657</a>] - Fixed an issue that caused GBM Grid Search to hang.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10675'>#10675</a>] - Subset h2o frame now allows 0 row subset - just as data.frame.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10712'>#10712</a>] - Fixed an issue that caused the R `apply` method to fail to work with `h2o.var()`.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10755'>#10755</a>] - PCA no longer reports errors when using PCA on wide datasets with `pca_method = Randomized`. Note that it is still not recommended to use Randomized with wide datasets.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10796'>#10796</a>] - Jenkins builds no longer all share the same R package directory, and new H2O R libraries are installed during testing. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10801'>#10801</a>] - When trimming is done, H2O now checks if it passes the beginning of the string. This check prevents the code from going further down the memory with negative indexes.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10866'>#10866</a>] - Stacked Ensembles no longer fails when the `fold_assignment` for base learners is not `Modulo`. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10881'>#10881</a>] - Fixed an issue that caused H2O to generate invalid code in POJO for PCA/SVM.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10969'>#10969</a>] - Instead of using random charset for getting bytes from strings, the source code now centralizes "byte extraction" in StringUtils. This prevents different build machines from using different default encoders.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10979'>#10979</a>] - When performing a Random Hyperparameter Search, if the model parameter seed is set to the default value but a search_criteria seed is not, then the model parameter seed will now be set to search_criteria seed+0, 1, 2, ..., model_number. Seeding the built models makes random hyperparameter searches more repeatable.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10989'>#10989</a>] - Fixed a bad link that was included in the "A K/V Store for In-Memory Analytics, Part 2" blog.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11027'>#11027</a>] - Comments are now permitted in Content-Type header for application/json mime type. As a result, specifying content-type charset no longer results in the request body being ignored. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11032'>#11032</a>] - Improved the Python `group_by` option count column name to match the R client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11035'>#11035</a>] - Fixed broken links in the "Hacking Algorithms into H2O" blog post.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11045'>#11045</a>] - The Python API now provides a method to extract parameters from `cluster_status`.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11060'>#11060</a>] - Fixed incorrect parsing of input parameters. Previously, system property parsing logic added the value of any system property other than "ga_opt_out" to the arguments list if a property was prefixed with "ai.h2o.". This caused an attempt to parse the value of a system property as if it were itself a system property and at times resulted in an "Unknown Argument" error. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11063'>#11063</a>] - Fixed intermittent pyunit_javapredict_dynamic_data_paramsDR. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11066'>#11066</a>] - Fixed orc parser test by setting timezone to local time.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11074'>#11074</a>] - H2O can now correctly handle preflight OPTIONS calls - specifically in the event of a (1) CORS request and (2) the request has a content type other than text/plain, application/x-www-form-urlencoded, or multipart/form-data.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11091'>#11091</a>] - In the REST API, POST of application/json requests no longer fails if requests expect required fields.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11105'>#11105</a>] - The R client `impute` function now checks for categorical values and returns an error if none exist.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11120'>#11120</a>] - Fixed a filepath issue that occurred on Windows 7 systems when specifying a network drive.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11123'>#11123</a>] - Added a response column to Stacked Ensembles so that it can be exposed in the Flow UI.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11124'>#11124</a>] - Updated the list of required packages on the H2O download page for the Python client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11139'>#11139</a>] - Updated the header in the Confusion Matrix to make the list of actual vs predicted values more clear.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11189'>#11189</a>] - Explicit 1-hot encoding in FrameUtils no longer generates an invalid order of column names. MissingLevel is now the last column.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11193'>#11193</a>] - Fixed an issue that caused ModelBuilder to leak xval frames if hyperparameter errors existed. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11200'>#11200</a>] - Fixed an issue that caused PCA model output to fail to display the Importance of Components.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11203'>#11203</a>] - When using the H2O Python client, the varimp() function can now be used in PCA to retrieve the Importance of Components details.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11204'>#11204</a>] - Fixed an issue that caused an ArrayIndexOutOfBoundsException in GLM.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11205'>#11205</a>] - When a main model is cloned to create the CV models, clearValidationMessages() is now called. Messages are no longer all thrown into a single bucket, which previously caused confusion with the `error_count()`.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11206'>#11206</a>] - ModelBuilder.message(...) now correctly bumps the error count when the message is an error.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11208'>#11208</a>] - Fixed an issue with unseen categorical levels handling in GLM scoring. Prediction with "skip" missing value handling in GLM with more than one variable no longer fails.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11210'>#11210</a>] - ModelMetricsRegression._mean_residual_deviance is now exposed. For all algorithms except GLM, this is the mean residual deviance. For GLM, this is the total residual deviance.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11216'>#11216</a>] - Fixed an issue that caused the`~` operator to fail when used in the Python client. Now, all logical operators set their results as Boolean.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11218'>#11218</a>] - Fixed an issue that caused an assertion error in GLM.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11220'>#11220</a>] - In GLM, fixed an issue that caused GLM to fail when `quasibinomial` was specified with a link other than the default. Specifying an incorrect link for the quasibinomial family will now result in an error message.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11239'>#11239</a>] - Improved the doc strings for `sample_rate_per_class` in R and Python.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11240'>#11240</a>] - Fixed a bug in the cosine distance formula.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11241'>#11241</a>] - Fixed an issue with CBSChunk set with long argument.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11252'>#11252</a>] - C0DChunk with con == NaN now works with strings.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11267'>#11267</a>] - When retrieving a Variable Importance plot using the H2O Python client, the default number of features shown is now 10 (or all if < 10 exist). Also reduced the top and bottom margins of the Y axis. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11269'>#11269</a>] - When retrieving a Variable Importance plot using the H2O R client, the default number of features shown is now 10 (or all if < 10 exist).</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11302'>#11302</a>] - Fixed an ORC stream parse.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11314'>#11314</a>] - Appended constant string to frame.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11379'>#11379</a>] - Fixed an issue with the View Log option in Flow.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11383'>#11383</a>] - The h2o.deepwater.available function is now working in the R API.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11426'>#11426</a>] - Fixed a bug with Log.info that resulted in bypassing log initialization.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11427'>#11427</a>] - LogsHandler now checks whether logging on specific level is enabled before accessing the particular log.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11430'>#11430</a>] - Fixed a logging issue that caused PID values to be set to an incorrect value. H2O now initializes PID before we initializing SELF_ADDRESS. This change was necessary because initialization of SELF_ADDRESS triggers buffered logged messages to be logged, and PID is part of the log header.</li>
</ul>
<h4>Epic</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10279'>#10279</a>] - Added supported for iSAX 2.0. This algorithm is a time series indexing strategy that reduces the dimensionality of a time series along the time axis. For example, if a time series had 1000 unique values with data across 500 rows, you can reduce this dataset to a time series that uses 100 unique values across 10 buckets along the time span. The following demos are available for more information:
<ul>
<li>Python - <a href='https://github.com/h2oai/h2o-3/blob/master/h2o-py/demos/isax2.ipynb'>https://github.com/h2oai/h2o-3/blob/master/h2o-py/demos/isax2.ipynb</a></li>
<li>R - <a href='https://github.com/h2oai/h2o-3/blob/master/h2o-r/demos/isax.R'>https://github.com/h2oai/h2o-3/blob/master/h2o-r/demos/isax.R</a></li>
</ul>
</li>
</ul>
<h4>New Feature</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13060'>#13060</a>] - Generate R bindings now available for REST API. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13119'>#13119</a>] - Flow: Implemented test infrastructure for Jenkins/CI.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13510'>#13510</a>] - The R client now reports to the user when memory limits have been exceeded.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13038'>#13038</a>] - Added support to impute missing elements for RandomForest.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15255'>#15255</a>] - Added a probability calibration plot function.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9479'>#9479</a>] - A new h2o.pivot() function is available to allow pivoting of tables.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10570'>#10570</a>] - MOJO support has been extended to K-Means models.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10737'>#10737</a>] - Added two new options in GBM and DRF: `calibrate_model` and `calibrate_frame`. These flags allow you to retrieve calibrated probabilities for binary classification problems. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10746'>#10746</a>] - In Stacked Ensembles, added support for passing in models instead of model IDs when using the R client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10863'>#10863</a>] - Added support for saving and loading binary Stacked Ensemble models.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10993'>#10993</a>] - Added support for idxmax, idxmin in Python H2OFrame to get an index of max/min values.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10994'>#10994</a>] - Added support for which.max, which.min support for R H2OFrame to get an index of max/min values.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11023'>#11023</a>] - A new h2o.sort() function is available in the H2O Python client. This returns a new Frame that is sorted by column(s) in ascending order. The column(s) to sort by can be either a single column name, a list of column names, or a list of column indices.</li>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11036'>#11036</a>] - Word2vec can now be used with the H2O Python client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11040'>#11040</a>] - Missing values are filled sequentially for time series data. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11057'>#11057</a>] - Enabled cors option flag behind the sys.ai.h2o. prefix for debugging.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11155'>#11155</a>] - Added support for converting a Word2vec model to a Frame. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11169'>#11169</a>] - Created a Capability rest end point that gives the client an overview of registered extensions.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11219'>#11219</a>] - When viewing a model in Flow, a new **Download Gen Model** button is available, allowing you to save the h2o-genmodel.jar file locally. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11311'>#11311</a>] - Added an `h2o.flow()` function to base H2O. This allows users to open up a Flow window from within R and Python.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11357'>#11357</a>] - The `parse_type` parameter is now case insensitive.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11363'>#11363</a>] - Added automatic reduction of categorical levels for Aggregator. This can be done by setting `categorical_encoding=EnumLimited`.</li>
<li>[NA] - In GBM and DRF, added two new categorical_encoding schemas: SortByResponse and LabelEncoding. More information about these options is available <a href="http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/algo-params/categorical_encoding.html">here</a>.
</li>
</ul>
<h4>Story</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10821'>#10821</a>] - Added support for Leave One Covariate Out (LOCO). This calculates row-wise variable importances by re-scoring a trained supervised model and measuring the impact of setting each variable to missing or its most central value (mean or median & mode for categoricals).</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10940'>#10940</a>] - Removed support for Java 6.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11163'>#11163</a>] - Integrated XGBoost with H2O core as a separate extension module.</li>
</ul>
<h4>Task </h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10420'>#10420</a>] - Users can now run predictions in R using a MOJO or POJO without running h2o running.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10977'>#10977</a>] - Created a test to verify that random grid search honors the `max_runtime_secs` parameter.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11084'>#11084</a>] - Removed javaMess.txt from scripts</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11127'>#11127</a>] - A new `node()` function is available for retrieving node information from an H2O Cluster.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11242'>#11242</a>] - Improved the R/Py doc strings for the `sample_rate_per_class` parameter.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11298'>#11298</a>] - Users can now optionally build h2o.jar with a visualization data server using the following: `./gradlew -PwithVisDataServer=true -PvisDataServerVersion=3.14.0 :h2o-assemblies:main:projects`</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11339'>#11339</a>] - Removed support for the following Hadoop platforms: CDH 5.2, CDH 5.3, and HDP 2.1.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11351'>#11351</a>] - Added the ability to go from String to Enum in PojoUtils.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11364'>#11364</a> - Continued modularization of H2O by removing reflections utils and replace them by SPI.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11366'>#11366</a>] - Removed the deprecated `h2o.importURL` function from the R API.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11375'>#11375</a>] - Stacked Ensembles now removes any unnecessary frames, vecs, and models that were produced when compiled.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11378'>#11378</a>] - Updated R and Python doc strings to indicate that users can save and load Stacked Ensemble binary models. In the User Guide, updated the FAQ that previously indicated users could not save and load stacked ensemble models.</li>
</ul>
<h4> Improvement </h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10011'>#10011</a>] - Improved error handling when users receive the following error:
`Error: lexical error: invalid char in json text.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10411'>#10411</a>] - In PCA, when the user specifies a value for k that is <=0, then all principal components will automatically be calculated.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10804'>#10804</a>] - Exposed metalearner and base model keys in R/Py StackedEnsemble object.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10962'>#10962</a>] - The `h2o.download_pojo()` function now accepts a `jar_name` parameter, allowing users to create custom names for the downloaded file. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10992'>#10992</a>] - Added port and ip details to the error logs for h2o cloud.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11030'>#11030</a>] - When using Hadoop with SSL Internode Security, the `-internal_security` flag is now deprecated in favor of the `-internal_security_conf` flag.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11058'>#11058</a>] - Scala version of udf now serializes properly in multinode.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11070'>#11070</a>] - Fixed an NPM warn message.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11073'>#11073</a>] - Updated the documentation for using H2O with Anaconda and included an end-to-end example.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11079'>#11079</a>] - Arguments in h2o.naiveBayes in R are now the same as Python/Java.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11096'>#11096</a>] - StackedEnsembles is now stable vs. experimental.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11145'>#11145</a>] - Introduced latest_stable_R and latest_stable_py links, making it easy to point users to the current stable version of H2O for Python and R.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11156'>#11156</a>] - In the R client, the default for `nthreads` is now -1. The documentation examples have been updated to reflect this change.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11196'>#11196</a>] - ModelMetrics can sort models by a different Frame.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11221'>#11221</a>] - The application type is now reported in YARN manager, and H2O now overrides the default MapReduce type to H2O type.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11305'>#11305</a>] - Added a title option to PrintMOJO utility</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11316'>#11316</a>] - Flow now uses ip:port for identifying the node as part of LogHandler.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11350'>#11350</a>] - Reduced the frequency of Hadoop heartbeat logging.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11369'>#11369</a>] - In GLM, quasibinomial models produce binomial metrics when scoring.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11376'>#11376</a>] - Implemented methods to get registered H2O capabilities in Python client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11377'>#11377</a>] - Implemented methods to get registered H2O capabilities in R client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11382'>#11382</a>] - Upgraded Flow to version 0.7.0</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11395'>#11395</a>] - Removed the `selection_strategy` argument from Stacked Ensembles.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11417'>#11417</a>] - In Stacked Ensembles, added support for passing in models instead of model IDs when using the Python client.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11420'>#11420</a>] - Provided a file that contains a list of licenses for each H2O dependency. This can be acquired using com.github.hierynomus.license.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11424'>#11424</a>] - H2O now explicitly checks if the port and baseport is within allowed port range.</li>
</ul>
<h4> Docs </h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9795'>#9795</a>] - Added documentation describing how to call Rapids expressions from Flow.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15440'>#15440</a>] - Added parameter descriptions for Naive Bayes parameter.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10838'>#10838</a>] - Added examples for Naive Bayes parameter.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10965'>#10965</a>] - Added `label_encoder` and `sort_by_response` to the list of available `categorical_encoding` options.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10984'>#10984</a>] - Added support for KMeans in MOJO documentation.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10968'>#10968</a>] - Added a topic to the Data Manipulation section describing the `group_by` function.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11029'>#11029</a>] - In the Productionizing H2O section of the User Guide, added an example showing how to read a MOJO as a resource from a jar file.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11071'>#11071</a>] - Improved the R and Python documentation for coef() and coef_norm().
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11072'>#11072</a>] - In the GLM section of the User Guide, added a topic describing how to extract coefficient table information. This new topic includes Python and R examples.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11073'>#11073</a>] - Added information about Anaconda support to the User Guide. Also included an IPython Notebook example.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11085'>#11085</a>] - Added Word2vec to list of supported algorithms on <a href='http://docs.h2o.ai/h2o/latest-stable/index.html#algorithms'>docs.h2o.ai</a>.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11090'>#11090</a>] - Uncluttered the H2O User Guide. Combined serveral topics on the left navigation/TOC. Some changes include the following:
<ul>
<li>Moved AWS, Azure, DSX, and Nimbix to a new Cloud Integration section.</li>
<li>Added a new **Getting Data into H2O** topic and moved the Supported File Formats and Data Sources topics into this.</li>
<li>Moved POJO/MOJO topic into the **Productionizing H2O** section.</li>
</ul>
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11095'>#11095</a>] - In the Security topic of the User Guide, added a section about using H2O with PAM authentication.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11100'>#11100</a>] - Documentation for `h2o.download_all_logs()` now informs the user that the supplied file name must include the .zip extension.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11107'>#11107</a> - Added an FAQ describing how to use third-party plotting libraries to plot metrics in the H2O Python client. This faq is available in the <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/faq/python.html'>FAQ > Python</a> topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11119'>#11119</a>] - Added an "Authentication Options" section to **Starting H2O > From the Command Line**. This section describes the options that can be set for all available supported authentication types. This section also includes flags for setting the newly supported Pluggable Authentication Module (PAM) authentication as well as Form Authentication and Session timeouts for H2O Flow.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11121'>#11121</a>] - Updated documentation to indicate that Word2vec is now supported for Python.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11142'>#11142</a>] - Added support for HDP 2.6 in the Hadoop Users section.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11147'>#11147</a>] - Added two FAQs within the GLM section describing why H2O's glm differs from R's glm and the steps to take to get the two to match. These FAQs are available in the <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/glm.html#faq'>GLM > FAQ</a> section.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11157'>#11157</a>] - Updated R examples in the User Guide to reflect that the default value for `nthreads` is now -1.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11170'>#11170</a>] - Updated the POJO Quick Start markdown file and Javadoc. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11179'>#11179</a>] - Added the `-principal` keyword to the list of Hadoop launch parameters.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11183'>#11183</a>] - In the Deep Learning topic, deleted the Algorithm section. The information included in that section has been moved into the Deep Learning FAQ.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11186'>#11186</a>] - Documented support for using H2O with Microsoft Azure Linux Data Science VM. Note that this is currently still a BETA feature.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11198'>#11198</a>] - Added an FAQ describing YARN resource usage. This FAQ is available in the <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/faq/hadoop.html'>FAQ > Hadoop</a> topic.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11226'>#11226</a>] - Added parameter descriptions for PCA parameters.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11227'>#11227</a>] - Added examples for PCA parameters.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11237'>#11237</a>] - A new h2o.sort() function is available in the H2O Python client. This returns a new Frame that is sorted by column(s) in ascending order. The column(s) to sort by can be either a single column name, a list of column names, or a list of column indices. Information about this function is available in the Python and R documentation.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11238'>#11238</a>] - Updated the "Using H2O with Microsoft Azure" topics. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11251'>#11251</a>] - Updated the "What is H2O" section in each booklet.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11274'>#11274</a>] - A Deep Water booklet is now available. A link to this booklet is on <a href='doc.h2o.ai'>docs.h2o.ai</a>.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11282'>#11282</a>] - Updated GLM documentation to indicate that GLM supports both multinomial and binomial handling of categorical values.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11283'>#11283</a>] - Added an FAQ describing the steps to take if a user encounters a "Server error - server 127.0.0.1 is unreachable at this moment" message. This FAQ is available in the <a href='http://docs.h2o.ai/h2o/latest-stable/h2o-docs/faq/r.html'>FAQ > R</a> topic.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11287'>#11287</a>] - Fixed documentation that described estimating in K-means. </li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11289'>#11289</a>] - Updated the documentation that described how to download a model in Flow.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11329'>#11329</a>] - The Data Sources topic, which describes that data can come from local file system, S3, HDFS, and JDBC, now also includes that data can be imported by specifying the URL of a file.</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11352'>#11352</a>] - H2O now supports GPUs. Updated the FAQ that indicated we do not, and added a pointer to Deep Water.</li>
</ul>
Ueno (3.10.4.8) - 5/21/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/8/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/8/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11013'>#11013</a>] - Python: Frame summary does not return Python object
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11204'>#11204</a>] - AIOOB with GLM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11220'>#11220</a>] - glm : quasi binomial with link other than default causes an h2o crash
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11222'>#11222</a>] - Create new /3/SteamMetrics REST API endpoint
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11321'>#11321</a>] - Steam hadoop user impersonation
</li>
</ul>
Ueno (3.10.4.7) - 5/8/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/7/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11278'>#11278</a>] - h2o on yarn: H2O does not respect the cloud name in case of flatfile mode
</li>
</ul>
Ueno (3.10.4.6) - 4/26/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/6/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11154'>#11154</a>] - Problem with h2o.uploadFile on Windows
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11230'>#11230</a>] - glm: get AIOOB exception on attached data
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11232'>#11232</a>] - External cluster always reports ""Timeout for confirmation exceeded!"
</li>
</ul>
Ueno (3.10.4.5) - 4/19/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/5/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11182'>#11182</a>] - Problem with h2o.merge in python
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11195'>#11195</a>] - Failing SVM parse
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11197'>#11197</a>] - Rollups computation errors sometimes get wrapped in a unhelpful exception and the original cause is hidden.
</li>
</ul>
Ueno (3.10.4.4) - 4/15/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/4/index.html</a>
<h4> Technical task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11133'>#11133</a>] - Add documentation on how to create a config file
</li>
</ul>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9741'>#9741</a>] - PCA Rotations not displayed in Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10971'>#10971</a>] - Sparse matrix cannot be converted to H2O
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11118'>#11118</a>] - Flow/Schema problem, predicting on frame without response returns empty model metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11135'>#11135</a>] - Proportion of variance in GLRM for single component has a value > 1
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11140'>#11140</a>] - HDP 2.6 add to the build
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11141'>#11141</a>] - Set timeout for read/write confirmation in ExternalFrameWriter/ExternalFrameReader
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11150'>#11150</a>] - GLM default solver gets AIIOB when run on dataset with 1 categorical variable and no intercept
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11174'>#11174</a>] - Correct exit status reporting ( when running on YARN )
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11176'>#11176</a>] - Documentation: Update GLM FAQ and missing_values_handling parameter regarding unseen categorical values
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11064'>#11064</a>] - H2O Flow UI Authentication
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11115'>#11115</a>] - Implement session timeout for Flow logins
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11178'>#11178</a>] - Document a new parameters for h2odriver.
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11069'>#11069</a>] - Wrap R examples in code so that they don&39;t run on Mac OS
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11104'>#11104</a>] - Export polygon function to fix CRAN note in h2o R package
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11137'>#11137</a>] - Add a parameter that ignores the config file reader when h2o.init() is called
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11128'>#11128</a>] - Extend Watchdog client extension so cluster is also stopped when the client doesn&39;t connect in specified timeout
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11177'>#11177</a>] - Set hadoop user from h2odriver
</li>
</ul>
Ueno (3.10.4.3) - 3/31/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/3/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10196'>#10196</a>] - ARFF parser parses attached file incorrectly
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10986'>#10986</a>] - Proxy warning message displays proxy with username and password.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11054'>#11054</a>] - h2o.import_sql_table works in R but on python gives error
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11056'>#11056</a>] - java.lang.IllegalArgumentException with PCA
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11076'>#11076</a>] - Impute does not handle catgoricals when values is specified
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11108'>#11108</a>] - Increase number of bins in partial plots
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11051'>#11051</a>] - h2o.transform can produce incorrect aggregated sentence embeddings
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10754'>#10754</a>] - Errors with PCA on wide data for pca_method = Power
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10991'>#10991</a>] - Introduce mode in which failure of H2O client ensures whole H2O clouds goes down
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11067'>#11067</a>] - Add support for IBM IOP 4.2
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11075'>#11075</a>] - Placeholder for: [SW-334]
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11080'>#11080</a>] - Remove minor version from hadoop distribution in buildinfo.json file
</li>
</ul>
Ueno (3.10.4.2) - 3/18/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11009'>#11009</a>] - Deep Learning: mini_batch_size >>> 1 causes OOM issues
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11024'>#11024</a>] - head(df) and tail(df) results in R are inconsistent for datetime columns
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11033'>#11033</a>] - GLM with family = multinomial, intercept=false, and weights or SkipMissing produces error
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11044'>#11044</a>] - glm hot fix: fix model.score0 for multinomial
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11022'>#11022</a>] - Add option to specify a port range for the Hadoop driver callback
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11028'>#11028</a>] - Support reading MOJO from a classpath resource
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10947'>#10947</a>] - Arff Parser doesn&39;t recognize spaces in attribute
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10988'>#10988</a>] - How to generate Precision Recall AUC (PRAUC) from the scala code
</li>
</ul>
<h4> Docs
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10870'>#10870</a>] - Documentation: Add documentation for word2vec
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11007'>#11007</a>] - Documentation: Add topic for using with IBM Data Science Experience
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/11038'>#11038</a>] - Document "driverportrange" option of H2O&39;s Hadoop driver
</li>
</ul>
Ueno (3.10.4.1) - 3/3/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-ueno/1/index.html</a>
<h4> Technical task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10837'>#10837</a>] - Documentation: Naive Bayes links to parameters section
</li>
</ul>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10714'>#10714</a>] - Error in predict, performance functions caused by fold_column
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10717'>#10717</a>] - Kmeans Centroid info not Rendered through Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10724'>#10724</a>] - PCA "Importance of Components" returns "data frame with 0 columns and 0 rows"
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10762'>#10762</a>] - Stratified sampling does not split minority class
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10781'>#10781</a>] - R Kmean&39;s user_point doesn&39;t get used
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10799'>#10799</a>] - Setting -context_path doesn&39;t change REST API path
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10826'>#10826</a>] - K-means Training Metrics do not match Prediction Metrics with same data
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10832'>#10832</a>] - h2o-py/tests/testdir_hdfs/pyunit_INTERNAL_HDFS_timestamp_date_orc.py failing
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10909'>#10909</a>] - gradle update broke the build
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10910'>#10910</a>] - H2O config (~/.h2oconfig) should allow user to specify username and password
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10923'>#10923</a>] - Flow/R/Python - H2O cloudInfo should show if cluster is secured or not
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10930'>#10930</a>] - FLOW fails to display custom models including Word2Vec
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10931'>#10931</a>] - Import json module as different alias in Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10932'>#10932</a>] - Stacked Ensemble docstring example is broken
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10933'>#10933</a>] - The autogen R bindings have an incorrect definition for the y argument
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10938'>#10938</a>] - AIOOB while training an H2OKMeansEstimator
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10955'>#10955</a>] - Fix bug in randomgridsearch and Fix intermittent pyunit_gbm_random_grid_large.py
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10956'>#10956</a>] - Typos in Stacked Ensemble Python H2O User Guide example code
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10963'>#10963</a>] - StackedEnsemble: stacking fails if combined with ignore_columns
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10973'>#10973</a>] - AIOOB in GLM
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10748'>#10748</a>] - Documentation: Add Data Munging topic for file name globbing
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10902'>#10902</a>] - Integration to add new top-level Plot menu to Flow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10929'>#10929</a>] - Add stddev to PDP computation
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10588'>#10588</a>] - Update h2o-py README
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10694'>#10694</a>] - Generate Python API tests for H2O Cluster commands
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10810'>#10810</a>] - Add documentation for python GroupBy class
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10811'>#10811</a>] - Document python&39;s Assembly and ConfusionMatrix classes, add python API tests as well
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10831'>#10831</a>] - Clean up R docs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10879'>#10879</a>] - Documentation: Summarize the method for estimating k in kmeans and add to docs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10899'>#10899</a>] - Update links to Stacking on docs.h2o.ai
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10912'>#10912</a>] - H2O config (~/.h2oconfig) should allow user to specify username and password
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10957'>#10957</a>] - Check if strict_version_check is TRUE when checking for config file
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10679'>#10679</a>] - Documentation: Add info about sparse data support
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10682'>#10682</a>] - h2o doc deeplearning: clarify what the (heuristics)defaults for auto are in categorical_encoding
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10815'>#10815</a>] - Saving/serializing currently existing, detailed model information
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10854'>#10854</a>] - Py/R: Remove unused &39;cluster_id&39; parameter
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10876'>#10876</a>] - Update GBM FAQ
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10887'>#10887</a>] - Documentation: Add info about imputing data in Flow and in Data Manipulation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10891'>#10891</a>] - Documentation: Add instructions for running demos
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10898'>#10898</a>] - AIOOB Exception with fold_column set with kmeans
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10946'>#10946</a>] - Modify h2oconnect function to accept config with connect_params field
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10950'>#10950</a>] - Change of h2o.connect(config) interface to support Steam
</li>
</ul>
Tverberg (3.10.3.5) - 2/16/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/5/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10744'>#10744</a>] - GLM with interaction parameter and cross-validation cause Exception
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10812'>#10812</a>] - pca: hangs on attached data
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10857'>#10857</a>] - StepOutOfRangeException when building GBM model
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10869'>#10869</a>] - py unique() returns frame of integers (since epoch) instead of frame of unique dates
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10872'>#10872</a>] - py date comparisons don&39;t work for rows > 1
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10873'>#10873</a>] - AstUnique drops column types
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10906'>#10906</a>] - In R, the confusion matrix at the end doesn’t say: vertical: actual, across: predicted
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10907'>#10907</a>] - AIOOB in GLM with hex.DataInfo.getCategoricalId(DataInfo.java:952) is the error with 2 fold cross validation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10927'>#10927</a>] - Parse fails when trying to parse large number of Parquet files
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/93'>private-#93</a>] - POJO doesn&39;t include Forest classes
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10935'>#10935</a>] - moment producing wrong dates
</li>
</ul>
Tverberg (3.10.3.4) - 2/3/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/4/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10858'>#10858</a>] - Importing data in python returns error - TypeError: expected string or bytes-like object
</li>
</ul>
Tverberg (3.10.3.3) - 2/2/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/3/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10732'>#10732</a>] - Standard Errors in GLM: calculating and showing specifically when called
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10882'>#10882</a>] - Decrease size of h2o.jar
</li>
</ul>
Tverberg (3.10.3.2) - 1/31/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li> Hotfix: Remove StackedEnsemble from Flow UI. Training is only supported from Python and R interfaces. Viewing is supported in the Flow UI.
</li>
</ul>
Tverberg (3.10.3.1) - 1/30/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tverberg/1/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15369'>#15369</a>] - Using asfactor() in Python client cannot allocate to a variable
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10034'>#10034</a>] - R API&39;s h2o.interaction() does not use destination_frame argument
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10597'>#10597</a>] - Errors with PCA on wide data for pca_method = GramSVD which is the default
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10639'>#10639</a>] - StackedEnsemble should work for regression
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10761'>#10761</a>] - h2o gbm : for an unseen categorical level, discrepancy in predictions when score using h2o vs pojo/mojo
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10779'>#10779</a>] - Negative indexing for H2OFrame is buggy in R API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10790'>#10790</a>] - Relational operators don&39;t work properly with time columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10859'>#10859</a>] - java.lang.AssertionError when using h2o.makeGLMModel
</li>
</ul>
<h4> Story
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10352'>#10352</a>] - StackedEnsemble: put ensemble creation into the back end
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13019'>#13019</a>] - Implement word2vec in h2o
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10540'>#10540</a>] - Ability to Select Columns for PDP computation in Flow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10777'>#10777</a>] - Add PCA Estimator documentation to Python API Docs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10798'>#10798</a>] - Documentation: Add information about Azure support to H2O User Guide (Beta)
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10251'>#10251</a>] - h2o.create_frame(): if randomize=True, `value` param cannot be used
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10353'>#10353</a>] - REST: implement simple ensemble generation API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10740'>#10740</a>] - Modify R REST API to always return binary data
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10741'>#10741</a>] - Safe GET calls for POJO/MOJO/genmodel
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10760'>#10760</a>] - Import files by pattern
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10780'>#10780</a>] - StackedEnsemble: Add to online documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10834'>#10834</a>] - Add Stacked Ensemble code examples to R docs
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10173'>#10173</a>] - Documentation: As a K-Means user, I want to be able to better understand the parameters
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10638'>#10638</a>] - StackedEnsemble: add tests in R and Python to ensure that a StackedEnsemble performs at least as well as the base_models
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10753'>#10753</a>] - Clean up the generated Python docs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10791'>#10791</a>] - Filter H2OFrame on pandas dates and time (python)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10808'>#10808</a>] - Provide way to specify context_path via Python/R h2o.init methods
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10827'>#10827</a>] - Modify gen_R.py for Stacked Ensemble
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10865'>#10865</a>] - Add Stacked Ensemble code examples to Python docstrings
</li>
</ul>
Tutte (3.10.2.2) - 1/12/2017
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tutte/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tutte/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10772'>#10772</a>] - Enable HDFS-like filesystems
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10713'>#10713</a>] - import functions required for r-release check
</li>
</ul>
Tutte (3.10.2.1) - 12/22/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tutte/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tutte/1/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10206'>#10206</a>] - Summary() doesn&39;t update stats values when asfactor() is applied
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10409'>#10409</a>] - rectangular assign to a categorical column does not work (should be possible to assign either an existing level, or a new one)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10523'>#10523</a>] - Numerical Column Names in H2O and R
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10593'>#10593</a>] - pred_noise_bandwidth parameter is not reproducible with seed
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10622'>#10622</a>] - Fix mktime() referencing from 0 base to 1 base for month and day
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10627'>#10627</a>] - Binary loss functions return error in GLRM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10644'>#10644</a>] - python hist() plotted bars overlap
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10647'>#10647</a>] - Python set_levels doesn&39;t change other methods
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10650'>#10650</a>] - h2o doc: glm grid search hyper parameters missing/incorrect listing. Presently glrm&39;s is marked as glm&39;s
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10662'>#10662</a>] - Partial Plot incorrectly calculates for constant categorical column
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10676'>#10676</a>] - h2o.proj_archetypes returns error if constant column is dropped in GLRM model
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10686'>#10686</a>] - GLRM loss by col produces error if constant columns are dropped
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10693'>#10693</a>] - isna() overwrites column names
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10709'>#10709</a>] - NullPointerException with Quantile GBM, cross validation, & sample_rate < 1
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10716'>#10716</a>] - R h2o.download_mojo broken - writes a 1 byte file
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10728'>#10728</a>] - Seed definition incorrect in R API for RF, GBM, GLM, NB
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10731'>#10731</a>] - h2o.glm: get AIOOB exception with xval and lambda search
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10393'>#10393</a>] - Supporting GLM binomial model to allow two arbitrary integer values
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10287'>#10287</a>] - Implement ISAX calculations per ISAX word
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10288'>#10288</a>] - Optimizations and final fixes for ISAX
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10568'>#10568</a>] - Implement GLM MOJO
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10412'>#10412</a>] - Variance metrics are missing from GLRM that are available in PCA
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10453'>#10453</a>] - py h2o.as_list() should not return headers
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10615'>#10615</a>] - Modify sum() calculation to work on rows or columns
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10636'>#10636</a>] - make sure that the generated R bindings work with StackedEnsemble
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10730'>#10730</a>] - Add HDP 2.5 Support
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9937'>#9937</a>] - Remove grid.sort_by method in Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10598'>#10598</a>] - Documentation: Add GLM to list of algorithms that support MOJOs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10689'>#10689</a>] - Documentation: Add quasibinomomial family in GLM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10580'>#10580</a>] - Add SLURM cluster documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10595'>#10595</a>] - Add memory check for GLRM before proceeding
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10663'>#10663</a>] - Check to make sure hinge loss works for GLRM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10700'>#10700</a>] - Add parameters from _upload_python_object to H2OFrame constructor
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10701'>#10701</a>] - Refer to .h2o.jar.env when detaching R package
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10702'>#10702</a>] - Call on proper port when exiting R/detaching package
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10703'>#10703</a>] - Modify search for config file in R api
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10715'>#10715</a>] - properly handle url in R docs from autogen
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10172'>#10172</a>] - Documentation: As a GLM user, I want to be able to better understand the parameters
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10655'>#10655</a>] - Fix bad/inconsistent/empty categorical (bitset) splits for DRF/GBM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10691'>#10691</a>] - Auto-generate R bindings
</li>
</ul>
Turnbull (3.10.1.2) - 12/14/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turnbull/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turnbull/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9735'>#9735</a>] - Starting h2o server from R ignores IP and port parameters
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10395'>#10395</a>] - Treat 1-element numeric list as acceptable when numeric input required
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10421'>#10421</a>] - h2o&39;s cor() breaks R&39;s native cor()
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10499'>#10499</a>] - h2o.get_grid isn&39;t working
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10512'>#10512</a>] - `cor` function should properly pass arguments
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10534'>#10534</a>] - Avoid confusing error message when column name is not found.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10536'>#10536</a>] - overwrite_with_best_model fails when using checkpoint
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10538'>#10538</a>] - plot.h2oModel in R no longer supports metrics with uppercase names (e.g. AUC)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10546'>#10546</a>] - Fix citibike R demo
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10600'>#10600</a>] - Create an Attribute for Number of Interal Trees in Python
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10604'>#10604</a>] - Error with early stopping and score_tree_interval on GBM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10634'>#10634</a>] - Python&39;s coef() and coef_norm() should use column name not index
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10654'>#10654</a>] - Perfbar does not work for hierarchical path passed via -h2o_context
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10385'>#10385</a>] - Show Partial Dependence Plots in Flow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10525'>#10525</a>] - Allow setting nthreads > 255.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10603'>#10603</a>] - Add RMSE, MAE, RMSLE, and lift_top_group as stopping metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10618'>#10618</a>] - Update h2o.mean in R to match Python API
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10486'>#10486</a>] - Document Partial Dependence Plot in Flow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10526'>#10526</a>] - Add R endpoint for cumsum, cumprod, cummin, and cummax
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10553'>#10553</a>] - Modify correlation matrix calculation to match R
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10561'>#10561</a>] - Remove max_confusion_matrix_size from booklets & py doc
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/128'>private-#128</a>] - aggregator should calculate domain for enum columns in aggregated output frames & member frames based on current output or member frame
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/115'>private-#115</a>] - Naive Bayes (and maybe GLM): Drop limit on classes that can be predicted (currently 1000)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10530'>#10530</a>] - Speed up GBM and DRF
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10653'>#10653</a>] - Support `-context_path` to change servlet path for REST API
</li>
</ul>
<h4> IT Help
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10194'>#10194</a>] - Adding a custom loss-function
</li>
</ul>
Turing (3.10.0.10) - 11/7/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turing/10/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turing/10/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10395'>#10395</a>] - Treat 1-element numeric list as acceptable when numeric input required
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10579'>#10579</a>] - Cannot determine file type
</li>
</ul>
Turing (3.10.0.9) - 10/25/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turing/9/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turing/9/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10458'>#10458</a>] - h2o.year() method does not return year
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10468'>#10468</a>] - Regression Training Metrics: Deviance and MAE were swapped
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10477'>#10477</a>] - h2o.max returns NaN even when na.rf condition is set to TRUE
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10500'>#10500</a>] - Fix display of array-valued entries in TwoDimTables such as grid search results
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10492'>#10492</a>] - Optimize algorithm for automatic estimation of K for K-Means
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/127'>private-#127</a>] - include flow, /3/ API accessible Aggregator model in h2o-3
</li>
</ul>
Turing (3.10.0.8) - 10/10/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turing/8/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turing/8/index.html</a>
<h4> Technical task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10275'>#10275</a>] - R binding for new MOJO
</li>
</ul>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10294'>#10294</a>] - S3 API method PersistS3uriToKey breaks expected contract
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10347'>#10347</a>] - GLM multinomial with defaults fails on attached dataset
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10351'>#10351</a>] - .structure() encounters list index out of bounds when nan is encountered in column
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10367'>#10367</a>] - max_active_predi tors option in glm does not work anymore
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10372'>#10372</a>] - Printed PCA model metrics in R is missing
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10388'>#10388</a>] - R - Unnecessary JDK requirement on Windows
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10416'>#10416</a>] - uuid columns with mostly missing values causes parse to fail.
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/171'>private-#171</a>] - Fold Column not available in h2o.grid
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14907'>#14907</a>] - Compute partial dependence data
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10332'>#10332</a>] - Create Method to Return Columns of Specific Type
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10402'>#10402</a>] - Find optimal number of clusters in K-Means
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10403'>#10403</a>] - Add optional categorical encoding schemes for GBM/DRF
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10242'>#10242</a>] - Tasks for completing MOJO support
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10356'>#10356</a>] - Ensure functions have `h2o.*` alias in R API
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10376'>#10376</a>] - Sync up functionality of download_mojo and download_pojo in R & Py
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10410'>#10410</a>] - Improve the stopping criterion for K-Means Lloyds iterations
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/174'>private-#174</a>] - Encryption of H2O communication channels
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/137'>private-#137</a>] - add option to Aggregator model to show ignored columns in output frame
</li>
</ul>
Turing (3.10.0.7) - 9/19/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turing/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turing/7/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10215'>#10215</a>] - NPE during categorical encoding with cross-validation (Windows 8 runit only??)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10221'>#10221</a>] - H2OFrame arithmetic/statistical functions return inconsistent types
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10230'>#10230</a>] - Multi file parse fails with NPE
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10285'>#10285</a>] - h2o.hist() does not respect breaks
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10311'>#10311</a>] - importFiles, with s3n, gives NullPointerException
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10319'>#10319</a>] - Python Structure() Breaks When Applied to Entire Dataframe
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9645'>#9645</a>] - Diff operation on column in H2O Frame
</li>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/154'>private-#154</a>] - calculate residuals in h2o-3 and in flow and create a new frame with a new column that contains the residuals
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9720'>#9720</a>] - Clean up Python booklet code in repo
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10211'>#10211</a>] - In R, allow x to be missing (meaning take all columns except y) for all supervised algo&39;s
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10244'>#10244</a>] - median() should return a list of medians from an entire frame
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10249'>#10249</a>] - Conduct rbind and cbind on multiple frames
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10297'>#10297</a>] - Add argument to H2OFrame.print in R to specify number of rows
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10328'>#10328</a>] - Suppress chunk summary in describe()
</li>
</ul>
Turing (3.10.0.6) - 8/25/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turing/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turing/6/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/private-h2o-3/issues/164'>private-#164</a>] - Hashmap in H2OIllegalArgumentException fails to deserialize & throws FATAL
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9809'>#9809</a>] - NPE in MetadataHandler
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10009'>#10009</a>] - hist() fails for constant numeric columns
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10095'>#10095</a>] - Client mode: flatfile requires list of all nodes, but a single entry node should be sufficient
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10126'>#10126</a>] - Make CreateFrame reproducible for categorical columns.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10127'>#10127</a>] - Fix intermittency of categorical encoding via eigenvector.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10130'>#10130</a>] - isBitIdentical is returning true for two Frames with different content
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10140'>#10140</a>] - AssertionError for DL train/valid with categorical encoding
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10154'>#10154</a>] - Wrong MAE for observation weights other than 1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10161'>#10161</a>] - H2ODriver for CDH5.7.0 does not accept memory settings
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10191'>#10191</a>] - H2OFrame.drop() leaves the frame in inconsistent state
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9932'>#9932</a>] - Implement skewness calculation for H2O Frames
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9933'>#9933</a>] - Implement kurtosis calculation for H2O Frames
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10051'>#10051</a>] - Add ability to do a deep copy in Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10086'>#10086</a>] - Add docs for h2o.make_metrics() for R and Python
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10136'>#10136</a>] - Add RMSLE to model metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10179'>#10179</a>] - Return unique values of a categorical column as a Pythonic list
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10152'>#10152</a>] - Refactor and simplify implementation of Pearson Correlation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10155'>#10155</a>] - Add MAE to CV Summary
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9640'>#9640</a>] - Create h2o.* functions for H2O primitives
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10021'>#10021</a>] - Add methods to get actual and default parameters of a model
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10055'>#10055</a>] - Add ability to drop a list of columns or a subset of rows from an H2OFrame
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10061'>#10061</a>] - Ensure all is*() functions return a list
</li>
</ul>
Turing (3.10.0.3) - 7/29/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turing/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turing/3/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9739'>#9739</a>] - Error when setting a string column to a single value in R/Py
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9894'>#9894</a>] - R h2o.merge() ignores by.x and by.y
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10058'>#10058</a>] - Download Logs broken URL from Flow
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9887'>#9887</a>] - H2O Version Check
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9947'>#9947</a>] - Add an h2o.concat function equivalent to pandas.concat
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9973'>#9973</a>] - Add Huber loss function for GBM and DL (for regression)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9994'>#9994</a>] - Add RMSE to model metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10027'>#10027</a>] - Add Mean Absolute Error to Model Metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10031'>#10031</a>] - Add mean absolute error to scoring history and model plotting
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10039'>#10039</a>] - Add categorical encoding schemes for DL and Aggregator
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10078'>#10078</a>] - Compute supervised ModelMetrics from predicted and actual values in Java/R
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10085'>#10085</a>] - Compute supervised ModelMetrics from predicted and actual values in Python
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14849'>#14849</a>] - Implement gradient checking for DL
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9567'>#9567</a>] - Add better warning message to functions of H2OModelMetrics objects
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9946'>#9946</a>] - Add demo datasets to Python package
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10036'>#10036</a>] - Replace "MSE" with "RMSE" in scoring history table
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10045'>#10045</a>] - Make all TwoDimTable Headers Pythonic in R and Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10052'>#10052</a>] - Achieve consistency between DL and GBM/RF scoring history in regression case
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10054'>#10054</a>] - Disable R^2 stopping criterion in tree model builders
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10072'>#10072</a>] - Remove R^2 from all model output except GLM
</li>
</ul>
Turin (3.8.3.4) - 7/15/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turin/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turin/4/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9963'>#9963</a>] - File parse from S3 extremely slow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10068'>#10068</a>] - Fix Deep Learning POJO for hidden dropout other than 0.5
</li>
</ul>
Turin (3.8.3.2) - 7/1/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turin/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turin/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13884'>#13884</a>] - DRF: sample_rate=1 not permitted unless validation is performed
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15026'>#15026</a>] - create a set of tests which create large POJOs for each algo and compiles them
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15228'>#15228</a>] - Merge (method="radix") bug1
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15231'>#15231</a>] - Merge (method="radix") bug2
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9506'>#9506</a>] - Fold Column not available in h2o.grid
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9893'>#9893</a>] - h2o.merge(,method="radix") failing 15/40 runs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9954'>#9954</a>] - Parse: java.lang.IllegalArgumentException: 0 > -2147483648
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9956'>#9956</a>] - Cached errors are not printed if H2O exits
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9995'>#9995</a>] - java.lang.ClassCastException for Quantile GBM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10000'>#10000</a>] - model_summary number of trees is too high for multinomial DRF/GBM models
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10002'>#10002</a>] - NPE when accessing invalid null Frame cache in a Frame&39;s vecs()
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10004'>#10004</a>] - TwoDimTable version of a Frame prints missing value (NA) as 0
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10012'>#10012</a>] - Fix tree split finding logic for some cases where min_rows wasn&39;t satisfied and the entire column was no longer considered even if there were allowed split points
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10016'>#10016</a>] - saveModel and loadModel don&39;t work with windows c:/ paths
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10018'>#10018</a>] - getStackTrace fails on NumberFormatException
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10019'>#10019</a>] - TwoDimTable for Frame Summaries doesn&39;t always show the full precision
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10020'>#10020</a>] - DRF OOB scoring isn&39;t using observation weights
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10022'>#10022</a>] - AIOOBE when calling &39;getModel&39; in Flow while a GLM model is training
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9619'>#9619</a>] - Properly document the addition of missing_values_handling arg to GLM
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14588'>#14588</a>] - Matt&39;s new merge (aka join) integrated into H2O
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9756'>#9756</a>] - Improved handling of missing values in tree models (training and testing)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9983'>#9983</a>] - IPv6 documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9989'>#9989</a>] - Stop GBM models once the effective learning rate drops below 1e-6.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/10017'>#10017</a>] - Log input parameters during boot of H2O
</li>
</ul>
Turchin (3.8.2.9) - 6/10/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/9/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/9/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9849'>#9849</a>] - Python apply() doesn&39;t recognize % (modulo) within lambda function
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9869'>#9869</a>] - Documentation: Add RoundRobin histogram_type to GBM/DRF
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9886'>#9886</a>] - Add "seed" option to GLM in documentation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9902'>#9902</a>] - Documentation: Update supported Hadoop versions
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9910'>#9910</a>] - Models hang when max_runtime_secs is too small
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9911'>#9911</a>] - Default min/max_mem_size to gigabytes in h2o.init
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9925'>#9925</a>] - Add "ignore_const_cols" argument to glm and gbm for Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9927'>#9927</a>] - AIOOBE in GBM if no nodes are split during tree building
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9929'>#9929</a>] - Negative R^2 (now NaN) can prevent early stopping
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9936'>#9936</a>] - Two grid sorting methods in Py API - only one works sometimes
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9680'>#9680</a>] - Add seed argument to GLM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9846'>#9846</a>] - Add cor() function to Rapids
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9930'>#9930</a>] - Verify checkpoint argument in h2o.gbm (for R)
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14989'>#14989</a>] - Sync up argument names in `h2o.init` between R and Python
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9924'>#9924</a>] - Change `getjar` to `get_jar` in h2o.download_pojo in R
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9926'>#9926</a>] - Change min_split_improvement default value from 0 to 1e-5 for GBM/DRF
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9938'>#9938</a>] - Allow specification of "AUC" or "auc" or "Auc" for stopping_metrics, sorting of grids, etc.
</li>
</ul>
Turchin (3.8.2.8) - 6/2/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/8/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/8/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9914'>#9914</a>] - Make Random grid search consistent between clients for same parameters
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9916'>#9916</a>] - Allow learn_rate_annealing to be passed to H2OGBMEstimator constructor in Python API
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9918'>#9918</a>] - Fix typo in GBM/DRF Python API for col_sample_rate_change_per_level - was misnamed and couldn&39;t be set
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9908'>#9908</a>] - Add a new metric: mean misclassification error for classification models
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9901'>#9901</a>] - No longer print negative R^2 values - show NaN instead
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9913'>#9913</a>] - Add xval=True/False as an option to model_performance() in Python API
</li>
</ul>
Turchin (3.8.2.6) - 5/24/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/6/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14860'>#14860</a>] - Number of active predictors is off by 1 when Intercept is included
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9871'>#9871</a>] - GLM with cross-validation AIOOBE (+ Grid-Search + Multinomial, may be related)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9872'>#9872</a>] - Improved accuracy for histogram_type="QuantilesGlobal" for DRF/GBM
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14669'>#14669</a>] - GLM needs &39;seed&39; argument for new (random) implementation of n-folds
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9680'>#9680</a>] - Add seed argument to GLM
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9857'>#9857</a>] - Remove _Dev from file name _DataScienceH2O-Dev
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9874'>#9874</a>] - Clean up overly long and duplicate error message in KeyV3
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9882'>#9882</a>] - Allow the user to pass column types of an existing H2OFrame during Parse/Upload in R and Python
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9883'>#9883</a>] - Tweak Parser Heuristic
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9884'>#9884</a>] - GLM improvements and fixes
</li>
</ul>
Turchin (3.8.2.5) - 5/19/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/5/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/5/index.html</a>
<h4> Technical task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9839'>#9839</a>] - Documentation update for relevel
</li>
</ul>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15189'>#15189</a>] - DRF: cannot compile pojo
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15210'>#15210</a>] - GBM pojo compile failures
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9808'>#9808</a>] - Bug in h2o-py H2OScaler.inverse_transform()
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9810'>#9810</a>] - Add NAOmit() to Rapids
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9827'>#9827</a>] - AIOOBE in Vec.factor (due to Parse bug?)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9833'>#9833</a>] - In grid search, max_runtime_secs without max_models hangs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9862'>#9862</a>] - GBM&39;s fold_assignment = "Stratified" breaks with missing values in response column
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9667'>#9667</a>] - Implement h2o.relevel, equivalent of base R&39;s relevel function
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9788'>#9788</a>] - Add Kerberos authentication to Flow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9823'>#9823</a>] - Summaries Fail in rdemo.citi.bike.small.R
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9825'>#9825</a>] - DimReduction for EasyModelAPI
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9844'>#9844</a>] - Make histograms truly adaptive (quantiles-based) for DRF/GBM
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9832'>#9832</a>] - Add a list of gridable parameters to the docs
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9834'>#9834</a>] - Add relevel() to Python API
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9835'>#9835</a>] - Improve the progress bar based on max_runtime_secs & max_models & actual work
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9838'>#9838</a>] - Improve GBM/DRF reproducibility for fixed parameters and hardware
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9841'>#9841</a>] - Check sanity of random grid search parameters (max_models and max_runtime_secs)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9842'>#9842</a>] - Add Job&39;s remaining time to Flow
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9848'>#9848</a>] - Add enum option &39;histogram_type&39; to DRF/GBM (and remove random_split_points)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9852'>#9852</a>] - JUnit: Separate POJO namespace during junit testing
</li>
</ul>
Turchin (3.8.2.3) - 4/25/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/3/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9783'>#9783</a>] - Incorrect sparse chunk getDoubles() extraction
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9759'>#9759</a>] - Create h2o.get_grid
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9767'>#9767</a>] - Implement distributed Aggregator for visualization
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9768'>#9768</a>] - Add col_sample_rate_change_per_level for GBM/DRF
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9769'>#9769</a>] - Add learn_rate_annealing for GBM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9770'>#9770</a>] - Add random cut points for histograms in DRF/GBM (ExtraTreesClassifier)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9782'>#9782</a>] - Add limit on max. leaf node contribution for GBM
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9780'>#9780</a>] - Add tests for early stopping logic (stopping_rounds > 0)
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9807'>#9807</a>] - Make NA split decisions internally more consistent
</li>
</ul>
Turchin (3.8.2.2) - 4/8/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9755'>#9755</a>] - Implement max_runtime_secs to limit total runtime of building GLM models with and without cross-validation enabled
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9749'>#9749</a>] - Add stratified sampling per-tree for DRF/GBM
</li>
</ul>
Turchin (3.8.2.1) - 4/7/2016
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turchin/1/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9701'>#9701</a>] - AIOOBE for quantile regression with stochastic GBM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9704'>#9704</a>] - Naive Bayes AIOOBE
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9706'>#9706</a>] - AIOOBE for GBM if test set has different number of classes than training set
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9709'>#9709</a>] - Number of CPUs incorrect in Flow when using a hypervisor
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9730'>#9730</a>] - Grid search runtime isn&39;t enforced for CV models
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9754'>#9754</a>] - AIOOBE in GLM for dense rows in sparse data
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9483'>#9483</a>] - Compute and display statistics of cross-validation model metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9708'>#9708</a>] - Add keep_cross_validation_fold_assignment and more CV accessors
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9710'>#9710</a>] - Set initial weights and biases for DL models
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9725'>#9725</a>] - Control min. relative squared error reduction for a node to split (DRF/GBM)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9740'>#9740</a>] - On-the-fly interactions for GLM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9749'>#9749</a>] - Add stratified sampling per-tree for DRF/GBM
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/15002'>#15002</a>] - Create test cases to show that POJO prediction behavior can be different than in-h2o-model prediction behavior
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/12992'>#12992</a>] - Populate start/end/duration time in milliseconds for all models
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9633'>#9633</a>] - Consistent handling of missing categories in GBM/DRF (and between H2O and POJO)
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9674'>#9674</a>] - Alert the user if columns can&39;t be histogrammed due to numerical extremities
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9693'>#9693</a>] - GLM should generate error if user enter an alpha value greater than 1.
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9698'>#9698</a>] - Create full holdout prediction frame for cross-validation predictions
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9703'>#9703</a>] - Support Validation Frame and Cross-Validation for Naive Bayes
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9744'>#9744</a>] - Add class_sampling_factors argument to DRF/GBM for R and Python APIs
</li>
</ul>
Turan (3.8.1.4) - 3/16/16
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turan/4/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turan/4/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13525'>#13525</a>] - KMeans: Size of clusters in Model Output is different from the labels generated on the training set
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14940'>#14940</a>] - GLM fails on negative alpha
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9656'>#9656</a>] - countmatches bug
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9665'>#9665</a>] - bug in processTables in communication.R
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9679'>#9679</a>] - Allow strings to be set to NA
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9657'>#9657</a>] - Implement Shannon entropy for a string
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9658'>#9658</a>] - Implement proportion of substrings that are valid English words
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9671'>#9671</a>] - Add utility function, h2o.ensemble_performance for ensemble and base learner metrics
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9678'>#9678</a>] - Add date/time and string columns to createFrame.
</li>
</ul>
<h4> Task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13073'>#13073</a>] - Certify sparkling water on CDH5.2
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13289'>#13289</a>] - Make python equivalent of as.h2o() work for numpy array and pandas arrays
</li>
</ul>
Turan (3.8.1.3) - 3/6/16
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turan/3/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turan/3/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9583'>#9583</a>] - Collinear columns cause NPE for P-values computation
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9659'>#9659</a>] - Update default values in h2o.glm.wrapper from -1 and NaN to NULL
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9660'>#9660</a>] - AIOOBE in NewChunk
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/13012'>#13012</a>] - Hive UDF form for Scoring Engine POJO for H2O Models
</li>
</ul>
Turan (3.8.1.2) - 3/4/16
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turan/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turan/2/index.html</a>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9651'>#9651</a>] - /3/scalaint fails with a 404
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9649'>#9649</a>] - Allow DL models to be pretrained on unlabeled data with an autoencoder
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9646'>#9646</a>] - H2O Flow does not contain CodeMirror library
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9648'>#9648</a>] - Model export fails: parent directory does not exist
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9650'>#9650</a>] - Flow doesn&39;t show DL AE error (MSE) plot
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9655'>#9655</a>] - Do not compute expensive quantiles during h2o.summary call
</li>
</ul>
Turan (3.8.1.1) - 3/3/16
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-turan/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-turan/1/index.html</a>
<h4> Technical task
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9643'>#9643</a>] - implement random (stochastic) hyperparameter search
</li>
</ul>
<h4> Bug
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9578'>#9578</a>] - Parse: Incorrect assertion error caused by very large few column data
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9588'>#9588</a>] - h2o::|,& operator handles NA&39;s differently than base::|,&
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9594'>#9594</a>] - h2o::as.logical behavior is different than base::as.logical
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9620'>#9620</a>] - Importing CSV file is not working with "java -jar h2o.jar -nthreads -1"
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9623'>#9623</a>] - Allow DL reproducible mode to work with user-given train_samples_per_iteration >= 0
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9628'>#9628</a>] - Grid Search NPE during Flow display after grid was cancelled
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9631'>#9631</a>] - NPE in initialMSE computation for GBM
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9634'>#9634</a>] - DL checkpoint restart doesn&39;t honor a change in stopping_rounds
</li>
</ul>
<h4> New Feature
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/14842'>#14842</a>] - Add option to train with mini-batch updates for DL
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9636'>#9636</a>] - Return leaf node assignments for DRF + GBM
</li>
</ul>
<h4> Improvement
</h4>
<ul>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9612'>#9612</a>] - Change default functionality of as_data_frame method in Py H2O
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9635'>#9635</a>] - Add method setNames for setting column names on H2O Frame
</li>
<li>[<a href='https://github.com/h2oai/h2o-3/issues/9641'>#9641</a>] - NPE in Log.write during cluster shutdown
</li>
</ul>
Tukey (3.8.0.6) - 2/23/16
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tukey/6/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tukey/6/index.html</a>
Enhancements
The following changes are improvements to existing features (which includes changed default values):
System
- [15269](https://github.com/h2oai/h2o-3/issues/15269): Handling Sparsity with Missing Values
- [9621](https://github.com/h2oai/h2o-3/issues/9621): Fix for erroneous conversion of NaNs to zeros during rebalancing
- [9622](https://github.com/h2oai/h2o-3/issues/9622): Remove bigdata test file (not available)
Bug Fixes
The following changes resolve incorrect software behavior:
Algorithms
- [9616](https://github.com/h2oai/h2o-3/issues/9616): CV models during grid search get overwritten
R
- [9587](https://github.com/h2oai/h2o-3/issues/9587): Di/trigamma handle NA
- [9617](https://github.com/h2oai/h2o-3/issues/9617): Progress bar for grid search with N-fold CV is wrong when max_models is given
Tukey (3.8.0.1) - 2/10/16
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tukey/1/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tukey/1/index.html</a>
New Features
These changes represent features that have been added since the previous release:
API
- [14762](https://github.com/h2oai/h2o-3/issues/14762): Ability to conduct a randomized grid search with optional limit of max. number of models or max. runtime
- [14781](https://github.com/h2oai/h2o-3/issues/14781): Add score_tree_interval to GBM to score every n'th tree
- [15217](https://github.com/h2oai/h2o-3/issues/15217): Make it easy for clients to sort by model metric of choice
- [9491](https://github.com/h2oai/h2o-3/issues/9491): Add ability to set a maximum runtime limit on all models
- [9571](https://github.com/h2oai/h2o-3/issues/9571): Return a grid search summary as a table with desired sort order and metric
Algorithms
- [private-665](https://github.com/h2oai/private-h2o-3/issues/665): Added ability to calculate GLM p-values for non-regularized models
- [13841](https://github.com/h2oai/h2o-3/issues/13841): Implemented gain/lift computation to allow using predicted data to evaluate the model performance
- [13143](https://github.com/h2oai/h2o-3/issues/13143): Compute the lift metric for binomial classification models
- [15123](https://github.com/h2oai/h2o-3/issues/15123): Add absolute loss (Laplace distribution) to GBM and Deep Learning
- [15308](https://github.com/h2oai/h2o-3/issues/15308): Add observations weights to quantile computation
- [15374](https://github.com/h2oai/h2o-3/issues/15374): For GBM/DRF, add ability to pick columns to sample from once per tree, instead of at every level
- [9535](https://github.com/h2oai/h2o-3/issues/9535): Quantile regression for GBM and Deep Learning
- [9565](https://github.com/h2oai/h2o-3/issues/9565): Add recall and specificity to default ROC metrics
Python
- [private-311](https://github.com/h2oai/private-h2o-3/issues/311): Added support for Python 3.5 and better (in addition to existing support for 2.7 and better)
Enhancements
The following changes are improvements to existing features (which includes changed default values):
Algorithms
- [15142](https://github.com/h2oai/h2o-3/issues/15142): Adjust string substitution and global string substitution to do in place updates on a string column.
Python
- [14844](https://github.com/h2oai/h2o-3/issues/14844): Fix layout issues of Python docs.
- [15241](https://github.com/h2oai/h2o-3/issues/15241): as.numeric for a string column only converts strings to ints rather than reals
- [15165](https://github.com/h2oai/h2o-3/issues/15165): Table printout in Python doesn't warn the user about truncation
- [15365](https://github.com/h2oai/h2o-3/issues/15365): Version mismatch message directs user to get a matching download
- [private-238](https://github.com/h2oai/private-h2o-3/issues/238): Implement secure Python h2o.init
- [15409](https://github.com/h2oai/h2o-3/issues/15409): Check and print a warning if a proxy environment variable is found
R
- [15241](https://github.com/h2oai/h2o-3/issues/15241): as.numeric for a string column only converts strings to ints rather than reals
- [15165](https://github.com/h2oai/h2o-3/issues/15165): Table printout in R doesn't warn the user about truncation
- [15335](https://github.com/h2oai/h2o-3/issues/15335): Improve R's reporting on quantiles
- [15365](https://github.com/h2oai/h2o-3/issues/15365): Version mismatch message directs user to get a matching download
Flow
- [15313](https://github.com/h2oai/h2o-3/issues/15313): Improve model convergence plots in Flow
- [9537](https://github.com/h2oai/h2o-3/issues/9537): Flow shows empty logloss box for regression models
- [9558](https://github.com/h2oai/h2o-3/issues/9558): Flow's histogram doesn't cover the full support
System
- [private-644](https://github.com/h2oai/private-h2o-3/issues/644): exportFile should be a real job and have a progress bar
- [15364](https://github.com/h2oai/h2o-3/issues/15364): Improve parse chunk size heuristic for better use of cores on small data sets
- [9547](https://github.com/h2oai/h2o-3/issues/9547): Print all columns to stdout for Hadoop jobs for easier debugging
Bug Fixes
The following changes resolve incorrect software behavior:
API
- [9572](https://github.com/h2oai/h2o-3/issues/9572): Ability to extend grid searches with more models
Algorithms
- [14826](https://github.com/h2oai/h2o-3/issues/14826): GLRM with Simplex Fails with Infinite Objective
- [15047](https://github.com/h2oai/h2o-3/issues/15047): Set GLM to give error when lower bound > upper bound in beta contraints
- [15101](https://github.com/h2oai/h2o-3/issues/15101): Set GLM to default to a value of rho = 0, if rho is not provided when beta constraints are used
- [15121](https://github.com/h2oai/h2o-3/issues/15121): Add check for epochs value when using checkpointing in deep learning
- [15150](https://github.com/h2oai/h2o-3/issues/15150): Set warnings about slowness from wide column counts comes before building a model, not after
- [15185](https://github.com/h2oai/h2o-3/issues/15185): Fix docstring reporting in iPython
- [15273](https://github.com/h2oai/h2o-3/issues/15273): Fix display of scoring speed for autoencoder
- [15332](https://github.com/h2oai/h2o-3/issues/15332): GLM gives different std. dev. and means than expected
- [9536](https://github.com/h2oai/h2o-3/issues/9536): Bad (perceived) quality of DL models during cross-validation due to internal weights handling
- [9566](https://github.com/h2oai/h2o-3/issues/9566): GLM with weights gives different answer h2o vs R
Python
- [15225](https://github.com/h2oai/h2o-3/issues/15225): sd not working inside group_by
- [15309](https://github.com/h2oai/h2o-3/issues/15309): Parser reads file of empty strings as 0 rows
- [15310](https://github.com/h2oai/h2o-3/issues/15310): Empty strings in Python objects parsed as missing
R
- [15225](https://github.com/h2oai/h2o-3/issues/15225): sd not working inside group_by
- [15140](https://github.com/h2oai/h2o-3/issues/15140): Fix bug in summary when zero-count categoricals were present.
- [14713](https://github.com/h2oai/h2o-3/issues/14713): Fix h2o.apply to correctly handle functions (so long as functions contain only H2O supported primitives)
System
- [14831](https://github.com/h2oai/h2o-3/issues/14831): Ability to ignore 0-byte files during parse
- [15307](https://github.com/h2oai/h2o-3/issues/15307): /Jobs fails if you build a Model and then overwrite it in the DKV with any other type
- [9544](https://github.com/h2oai/h2o-3/issues/9544): Improve progress bar for grid/hyper-param searches
---
Tibshirani (3.6.0.9) - 12/7/15
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tibshirani/9/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tibshirani/9/index.html</a>
New Features
These changes represent features that have been added since the previous release:
API
- [15100](https://github.com/h2oai/h2o-3/issues/15100): H2O now allows selection of the `non_negative` flag in GLM for R and Python
Algorithms
- [PUBDEB-1540](https://github.com/h2oai/h2o-3/issues/14512): Added Generalized Low-Rank Model (GLRM) algorithm
- [13068](https://github.com/h2oai/h2o-3/issues/13068): Added gains/lift computation
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/f4d2f85f3e61cfa03a1c80ea3a973e9d2f3ba20b): Added `remove_colinear_columns` parameter to GLM
R
- [15019](https://github.com/h2oai/h2o-3/issues/15019): R now retrieves column types for a H2O Frame more efficiently
Python
- [15200](https://github.com/h2oai/h2o-3/issues/15200): Added Python equivalent for `h2o.num_iterations`
- [15142](https://github.com/h2oai/h2o-3/issues/15142): Added `sub` and `gsub` to Python client
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/dbef536dc48fa56eb21d65339f1f1857724012f1): Added weighted quantiles to Python API
- [14280](https://github.com/h2oai/h2o-3/issues/14280): Added `sapply` operator to Python
- [14933](https://github.com/h2oai/h2o-3/issues/14933): H2O now plots decision boundaries for classifiers in Python
Enhancements
The following changes are improvements to existing features (which includes changed default values):
Algorithms
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/315201eb41a8439a1a77fb632b0e32a8181a2785): Change in behavior in GLM beta constraints - when ignoring constant/bad columns, remove them from `beta_constraints` as well
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/8035f7388560801e4af69c92c81a436a96f02957): Added `ignore_const_cols` to all algos
- [15217](https://github.com/h2oai/h2o-3/issues/15217): Improved ability to sort by model metric of choice in client
Python
- [15315](https://github.com/h2oai/h2o-3/issues/15315): H2O now checks for `H2O_DISABLE_STRICT_VERSION_CHECK` env variable in Python [GitHub commit](https://github.com/h2oai/h2o-3/commit/238a1c80e091c681da8b4b52eca308281244a7b9)
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/953b892560242146af0ffc124f32e7a9ddb89b04): H2O now allows l/r values to be null or an empty string
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/7e368a4df9c39b30ca4f4eac77e2c739ef38bbf6): H2O now accomodates `LOAD_FAST` and `LOAD_GLOBAL` in `bytecode_to_ast`
R
- [14352](https://github.com/h2oai/h2o-3/issues/14352): In R, `h2o.getTimezone()` previously returned a list of one, now it just returns the string
System
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/e1c9272bc21d4fcf3ab55f4fe646656a061ce542): Added more tweaks to help various low-memory configurations
Bug Fixes
The following changes resolve incorrect software behavior:
API
- [14991](https://github.com/h2oai/h2o-3/issues/14991): `h2o.grid` failed when REST API version was not default
- [15307](https://github.com/h2oai/h2o-3/issues/15307): `/Jobs` failed if you built a Model and then overwrote it in the DKV with any other type [GitHub commit](https://github.com/h2oai/h2o-3/commit/453db68b4d31130bdaa2b6a31cf80a88447baabf)
- [15299](https://github.com/h2oai/h2o-3/issues/15299): `/3/Jobs` failed with exception after running `/3/SplitFrame`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/888cc021694fe0b34aadba70bdd0d79811cfb30d): #15332 - Fixed error where sd and mean were adjusted to weights even if no observation weights were passed
Algorithms
- [15303](https://github.com/h2oai/h2o-3/issues/15303): GLRM validation frames must have the same number of rows as the training frame
- [15000](https://github.com/h2oai/h2o-3/issues/15000): Fixed assertion failure in Deep Learning
- [15221](https://github.com/h2oai/h2o-3/issues/15221): Could not compile POJO using K-means
- [15223](https://github.com/h2oai/h2o-3/issues/15223): Could not compile POJO using PCA
- [15226](https://github.com/h2oai/h2o-3/issues/15226): Could not compile POJO using Naive Bayes
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/928002f60a4523ad1b065635f55ccd2bb2b9091c): Fixed weighted mean and standard deviation computation in GLM
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/147fd0990b484747439ba0921f72763211d10072): Fixed stopping criteria for lambda search and multinomial in GLM
Python
- [15170](https://github.com/h2oai/h2o-3/issues/15170): H2OFrame indexing was no longer Pythonic on Bleeding Edge 10/23
- [15185](https://github.com/h2oai/h2o-3/issues/15185): Trying to get help in python client displayed the frame
- [15278](https://github.com/h2oai/h2o-3/issues/15278): Fixed ASTEQ `str_op` bug [GitHub commit](https://github.com/h2oai/h2o-3/commit/4fc88caa8d86679c4ed584d6a2c32ec931f03cdd)
R
- [14713](https://github.com/h2oai/h2o-3/issues/14713): `h2o.apply` did not correctly handle functions
- [15241](https://github.com/h2oai/h2o-3/issues/15241): R: as.numeric for a string column only converted strings to ints rather than reals
- [15225](https://github.com/h2oai/h2o-3/issues/15225): R: `sd` was not working inside `group_by`
- [15304](https://github.com/h2oai/h2o-3/issues/15304): R: Ignore Constant Columns was not an argument in Algos in R like it is in Flow
- [15053](https://github.com/h2oai/h2o-3/issues/15053): When a dataset was sliced, the int mapping of enums was returned
- [15314](https://github.com/h2oai/h2o-3/issues/15314): Improved handling when H2O has already been shutdown in R [GitHub commit](https://github.com/h2oai/h2o-3/commit/00d1505681135960722bc74def2a951ba01682ba)
- [15140](https://github.com/h2oai/h2o-3/issues/15140): Fixed categorical levels mapping bug
System
- [15309](https://github.com/h2oai/h2o-3/issues/15309): Parser read file of empty strings as 0 rows [GitHub commit](https://github.com/h2oai/h2o-3/commit/eee101ba16e11bcbc10e35ddd1f96bfb749156d3)
- [15310](https://github.com/h2oai/h2o-3/issues/15310): Empty strings in python objects were parsed as missing [GitHub commit](https://github.com/h2oai/h2o-3/commit/fb4817a6847192a9dc3f359b0f3ec2d63268aaef)
- [15282](https://github.com/h2oai/h2o-3/issues/15282): Save Model (Deeplearning): the filename for the model metrics file is too long for windows to handle
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/76aaf1ce863ea3cdfc5531dc0b551a26ff3ddf48): Fixed streaming load bug for large files
- [15150](https://github.com/h2oai/h2o-3/issues/15150): Column width slowness warning now prints before model build, not after
---
Tibshirani (3.6.0.7) - 11/23/15
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tibshirani/7/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tibshirani/7/index.html</a>
Enhancements
The following changes are improvements to existing features (which includes changed default values):
Algorithms
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/47f4b754e8f4178b86d5c4f929a46223b3f7b042): Added Iterations and Epochs to DL job status updates, added Iterations to scoring history
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/aa25d5f71b281): Cleaned up iteration counter to work for checkpointing
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/1038cdf9372b799d): Cleaned up counter iteration logic
Bug Fixes
The following changes resolve incorrect software behavior:
Algorithms
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/bd784804b5ba09fd21eb0ee67d13925b668496a3): Fixed scoring speed display for autoencoder, was showing 0 because wrong runtime was used (ms since 1970 instead of actual runtime)
---
Tibshirani (3.6.0.2) - 11/5/15
Download at: <a href='http://h2o-release.s3.amazonaws.com/h2o/rel-tibshirani/2/index.html'>http://h2o-release.s3.amazonaws.com/h2o/rel-tibshirani/2/index.html</a>
New Features
Algorithms
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/93cd9246762cab74008ae188e9dcad552978875a): Added support for grid search
- [15180](https://github.com/h2oai/h2o-3/issues/15180): Implemented GLRM grid search in R and Python
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/ada571ee5bfc7d8e140ab3fd3d416a8246909201): #15196: Enabled early convergence-based stopping by default for Deep Learning
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/514aa26de192821f7f5c780caed561f2689761ab): Added L1+LBFGS solver for multinomial GLM
Python
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/33fdf12410bbe0fc294a848a6b37719fa05e6f9a): #15196: Added Python API for convergence-based stopping
R
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/1f9dbffbea6f648717889064a9b612ee0fbbd3ab): Added `.Last` to `Delete InitID`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/723d552f7f1a9977c6fc837704b73cf1b85f0524): #15196: Enabled convergence-based early stopping for R API of Deep Learning
Enhancements
Algorithms
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/6e0c3575c2d5f07fcc9dacf51f76ec83279d1783): Enable grid search for Deep Learning parameters `overwrite_with_best_model`, `momentum_ramp`, `elastic_averaging`, `elastic_averaging_moving_rate`, & `elastic_averaging_regularization`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/e69a8a5da0a738b1fa46fac5614171269e042f5a): #15196: Stopping tolerance and stopping metric are no longer hidden if `stopping_rounds` is 0
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/514adb7bd456bc2d0ed08eb62db450de3524b935): Added checks to verify the mean, median, nrow, var, and sd are calculated correctly in `groupby`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/c2095f7ae5fafcb7e1bcc88c314005886725aba8): `mean` and `sd` now return lists
Python
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/3f0d0cb230aa047295524043dae46cc9cb19c7c5): [#15165] H2O now gives users [row x col] of Frame in `__str__`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/cde0510be3adc418b3595989761f3f5cdf513886): `sd`/`var` is now sampled for `group_by`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/30743335bf914af11d787bf9ec1961ac51ebde14): Parameter checking is now split between float and strings/unicode
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/ba630573dbf4cb571302710698664f4fda51f0b6): H2O now only wipes `src._ex` if `src_in_self`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/ebc53e6860028be1bcf220c9898863da0f915943): Refactored default arg handling in `astfun`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/e9950c6ea7abaac46c9670ba02016aef529a14ec): Added new parameters to estimators
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/fbfac8c6ed86cafe4b5f0b78f83e9c9c4e4221f4): Added session start/end; Python now ends the session on exit
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/c73b988ebf30168ddb6c76b6ed51a3c011b6504a): `src` and `self` types are now checked for `None`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/94a90ea8b7b859f2909a0eadf37b5949e722f2ea): H2O now passes caches through all prefix ops
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/edd0c7feffb32249a1ba0e596685483e81bfd34a): H2O now pushes cached types, names, and ncols forward if possible
R
- [14915](https://github.com/h2oai/h2o-3/issues/14915): Removed the R backward compatibility shim
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/c7652e8c93384b1be38fb88da18e65f954b6373c): Added [rows x cols] to `print.Frame` in R
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/a82297bd1f6c0e08722ac6671797fa0e41db77a8): `sd` can now alias `sdev` in `group_by`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/5c9bd93ec15f65658809def2634b13bd788ee7a1): Changed `.eval.driver` to `.fetch.data` in `h2o.getFrame`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/e9608bfc696e1a64d9becf865571fe11da21ac93): Removed debug printing of `==Finalizer on` in R
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/d6bc387a3d0d5f830f053a8384062eccb8b96d74): Added metalearning function
System
- [private-280](https://github.com/h2oai/private-h2o-3/issues/280): Added EasyPOJO comments and improvements
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/ba75b35fbc7498b3d0d3c0b1d362a44dba6f31bb): [#15115] Enabled `VectoCategoricalVec` to convert string columns to categorical columns
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/613e121e9ab13f32df4beffce1c2139607b5f64d): `apply` now works in
Bug Fixes
Algorithms
- [15223](https://github.com/h2oai/h2o-3/issues/15223): PCA: Could not compile POJO
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/545c00f9913b0fc586490ff391759f106228c121): [#15223] Incorrect PCA code was generated
Python
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/fc7e364c6c4cab889cbb5c8cab223ab092a301c5): #15203: Python was not updating exception on job update
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/b5f81871265d5c3cab4788313a719460dff32dbc): Added missing arguments to DRF/GBM/DL in scikit-learn-like API
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/fbe8275e1aec2ca930373fc18d65b11e706b88b5): Fixed `impute` in Python
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/ce343692ce2382e789959e717bf7a3996b5d9fda): Restored `ASTRename`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/f55dc318d74352333c7f2ec4f54fffad3bb49529): Fixed reference to `_quoted` in H2O module
R
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/cc34b88ff322ec829f3d52dca0d03b2c658dd6d6): [#15207, 15207] Hidden grid parameter was passed incorrectly from R
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/1ac19955ca1b0ce72e794542d58b4625affae525): H2O now uses deep copy when using `assign` from one global to another
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/a0300b0b519d06707598a31cf8aaf168835f869f): Fixed `getFrame` and directory `unlink`
System
- [14783](https://github.com/h2oai/h2o-3/issues/14783): `h2o.init()` failed to launch on the Docker image
- [14992](https://github.com/h2oai/h2o-3/issues/14992): Deep Learning generated an assertion error
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/c3544016927d93695bed20969d8588f159b5797d): Fixed rm handling of non-frames
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/2285a7c89effd9c9be8d45b5899c7463c335006f): Fixed `log_level`
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/f0f3559bc00e9f1ef264138db75b3beade32d066): Fixed eq2 slot assign
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/150a29fce82f1e0c1450b0c653342fbdfabf326f): Fixed a bug found during benchmarking for small data
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/3097a0754b6f49a7358a5ea3a8150606f36d5172): #15201: User-given weights were accidentally passed to N-fold CV models
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/91d141f5dd574293ecce6336365ec56fb8007a2d): Fixed NPE in Grid Schema
- [GitHub commit](https://github.com/h2oai/h2o-3/commit/41bf4a3c164734d1dd6d6ca010b1a3eac313cfdc): #15196: Convergence checks are now numerically stable
---
Scan your Python project for dependency vulnerabilities in two minutes
Scan your application