0.7.4

Enhancements

• #184 Change Keras+TF Neural Networks to scikit-learn MLP
• #233 Limit staking number of classes and models
• #232 Remove Linear model from Compete mode
• #208 Improve importance computation for large number of columns
• #205 Remove small learning rates for Xgboost

Bug fixes:

• #231 Restricted characters in feature_neams in Xgboost
• #227 Fix strings in golden_features.json - thank you @SuryaThiru!
• #215 Assure at least 20 samples (or k_folds) for each class

Docs update:

• #213 Update docs in AutoML - thank you @shahules786!

0.7.3

New features ✨

• #176 extended EDA - thanks to @shahules786

Bug fixes 🐛

• #201 error in golden features sampling
• #199 bug for float multi-class labels
• #196 add exception for empty data
• #195 set threshold for accuracy metric instead f1
• #194 ensemble should be best model if has more than 1 model
• #193 fixed predict aflter model loading
• #192 update pyarrow
• #191 hide shap warnings
• #190 fix in preprocessing
• #188 fix type in feature selection - thanks to @uditswaroopa

0.7.2

Bug fixes 🐛

• #187 fix wrong order in golden features step
• #186 fix _get_results_path
• #185 fix models loading
• #184 exception when drop all features during selection
• #182 catch exceptions from model and log to errors.md
• #181 remove forbidden characters in EDA
• #177 change docstring to google-stype
• #175 remove tuning_mode parameter from AutoML

0.7.1

Bug fixes 🐛

• #173 fix bug in shap sampling
• #174 update dtreeviz package

0.7.0

Improvements

• (#148) make AutoML scikit-learn compatible, thank you @spamz23! 👏 👏 👏
• (#170, #171 ) improve printouts while training AutoML

0.6.1

Enhancements

• #145 Add EDA for input data set #125
• #135 Add ability to pause and restore the training
• #19 Add tests for ensemble save and load

Refactor

• #149 Add Time Controller
• #80 add tests for one column input

Bug fixes

• #144 AutoMlException
• #142 Error when training NN on BNP Paribas kaggle dataset

0.6.0

• Add golden features transformer (#126)
• Add feature selection (#133)
• Add one-hot encoding (#76)
• Fixes in Neural Networks (#131, #129)
• Add max_depth in Extra Trees and Random Forest (#106)
• Add support for date/time features (#122)

0.5.5

• Speed-up CatBoost by removing very small learning rates
• Set larger tolerance in Logistic Regression, from 1e-4 to 5e-4

0.5.4

• #129 fix NN bug
• #128 add support for text features
• #122 add support for date/time features
• #106 tune max_depth in Random Fores and Extra Trees

0.5.3

(#122) Add support for date/time features