This library aims at providing tools for an automatic machine learning approach. As many tools already exist to establish one or the other component of an AutoML approach, the idea of this library is to provide a structure rather than to implement a complete service.
Three modifications:
1) The predictions are now also stored in private attributes as dataframes, and can be accessed with the getters: "predict_df", "predict_proba_df", "val_predict_df", "val_predict_proba_df".
2) Implementation of predict and predict_proba methods to test new data. These functions return either array as done by sklearn predict and sklearn predict_proba. They can also return dataframes if argument return_df is set to True (default to False). These functions overwrite the previously listed dataframes.
3) 3 tests added to test the correct functionning of these features.
Three modifications: 1) The predictions are now also stored in private attributes as dataframes, and can be accessed with the getters: "predict_df", "predict_proba_df", "val_predict_df", "val_predict_proba_df". 2) Implementation of predict and predict_proba methods to test new data. These functions return either array as done by sklearn predict and sklearn predict_proba. They can also return dataframes if argument return_df is set to True (default to False). These functions overwrite the previously listed dataframes. 3) 3 tests added to test the correct functionning of these features.