tabensemb.data.dataprocessor.RFEFeatureSelector#

class tabensemb.data.dataprocessor.RFEFeatureSelector(**kwargs)[source]#

Bases: AbstractFeatureSelector

Select features using recursive feature elimination, adapted from the implementation of RFECV in sklearn. Available arguments:

n_estimators: int

The number of trees used in random forests.

step: int

The number of eliminated features at each step.

min_features_to_select: int

The minimum number of features.

method: str

The method of calculating importance. “auto” for default impurity-based method implemented in RandomForestRegressor, and “shap” for SHAP value (which may slow down the program but is more accurate).

Methods

__init__(**kwargs)#

_defaults()

Defaults values for arguments defined in _cls_required_kwargs() and _required_kwargs()

_get_feature_names_out(data, datamodule)

Get selected features.