evaluate_features_to_target

evaluate_features_to_target#

mlproject.corr_analysis.dependency_graph.evaluate_features_to_target(X_source, y_target, n_splits=5, scoring=None)[source]#

Predict a target from features using KFold cross-validation.

Parameters:

  • X_source (pd.DataFrame) – Feature matrix.

  • y_target (pd.Series or np.array) – Target values.

  • n_splits (int) – Number of CV splits.

  • scoring (dict) – Dictionary of scoring functions for cross_validate.

Returns:

Per-fold metrics. summary_df (pd.DataFrame): Mean ± std of metrics across folds.

Return type:

metrics_df (pd.DataFrame)

Contents