Mljar-supervised

Enhancements
- 81 add scatter plot predicted vs target in regression
- 158 add ROC curve for binary classification
- 336 add visualization for Optuna results
- 352 add support for Colab
- 374 update seaborn
- 378 set golden features number
- 379 switch off boost_on_errors step in Optuna mode
- 380 add custom cross validation strategy
- 386 add correlation heatmap
- 387 add residual plot
- 389 add feature importance heatmap
- 390 add custom eval metric
- 393 update sklearn

Bug fixes
- 308 fix error in kaggle kernel
- 353, 355, 366, 368, 376, 382, 383, 384 fixes

Docs
- 391 add info about hyperparameters optimization methods

Big thank you for help for: ecoskian, xuzhang5788, xiaobo, RafaD5, drorhilman, strelzoff-erdc, muxuezi, tresoldi THANK YOU !!!

Enhancements
- 343 set seed in Optuna
- 344 set eval_metric directly in all algorithms
- 350 add estimated train time in Optuna mode
- 342 add `optuna_verbose` param in `AutoML()`
- 354 add KNN in Optuna
- 356 and Neural Network in Optuna
- 357, 348 use mljar wrapper for Random Forest and Extra Trees
- 358 add `extra_tree` param in LightGBM
- 359 switch off feature engineering in Optuna mode - only highly tuned models are produced
- 361 list all `eval_metric` in error message
- 362 add accuracy `eval_metric`
- 340 support for r2

Bug fixes
- 347 dont include Optuna tuning time in `total_time_limit`
- 360 missing auc scores for training in CatBoost

Add support to Python 3.9 (339) Thanks to rterbush!

Enhancements

- 332 We added Optuna framework for hyperparameters tuning. It can be used by setting `mode="Optuna"` in AutoML. You can read more details at blog post: https://mljar.com/blog/automl-optuna/

Enhancements
- 179 add `need_retrain()` method to detect performance decrease
- 226 extract rules from decision tree
- 310 add support for MAPE
- 312 optimize prediction time
- 313 set stacking time threshold depending on best model train time
- 320 search for model with prediction time constraint
- 322 `n_jobs` as a parameter
- 328 disable stacking for small (nrows < 500) datasets

Bug fixes
- 214 move directory after training
- 246 raise exception when small time limit and no models are trained
- 247 proper display for optimize AUC and R2
- 306 add `mix_encoding` argument in `AutoML` constructor
- 308 fix dependencies error in kaggle notebook
- 314 bug fix in hill climbing in Perform mode
- 323 fix catboost bug with tree limit
- 324 325 bug for feature importance for small data