pyreal.RealApp.prepare_feature_importance#

RealApp.prepare_feature_importance(x_train_orig=None, y_train=None, model_id=None, algorithm=None, shap_type=None, training_size=None)[source]#

Initialize and fit a global feature importance explainer

Parameters:

  • model_id (int or string) – Model id to explain

  • x_train_orig (DataFrame of shape (n_instances, n_features)) – Training data, if not provided at initialization.

  • y_train (DataFrame or Series) – Training targets, if not provided at initialization

  • algorithm (string) – GFI algorithm to use

  • shap_type (string) – If algorithm is “shap”, type of shap to use

  • training_size (int) – Number of rows to use in fitting explainer

Returns:

A fit GlobalFeatureImportance explainer