Changelog:
- Add `fit` API call for the studies. Returns a trained model directly.
- Random states bugfixing for studies, explainers etc.
- Add `feature_selection` parameters to the studies.
- sklearn params bugfixing: RF, bagging, adaboost, version 1.2
- update default imputers: [ice, missforest, hyperimpute, mean]
- data_cleanup plugin: drop constant features, high VIF features(multicollinearity). Added by default before the prediction layer.
- Data Encoding bugfixing: handle missing/unseen values.
- update env variables:
- `N_LEARNERS_JOBS` : number of CPUs to use by base learners
- `N_OPT_JOBS`: number of CPUs to use for hyperparam search
- `REDIS_HOST` : Redis IP address. useful when running in distributed clusters.
- `REDIS_PORT` : Redis port.
- Update HyperImpute
- Add evaluation tools for multiple seeds.
Tutorials:
- add MICE tutorial
- add categorical imputation tutorial
- add serialization tutorial