holisticai.explainability.metrics.rank_alignment

holisticai.explainability.metrics.rank_alignment(conditional_feature_importance: ConditionalImportances, ranked_feature_importance: Importances, aggregation=True)

Compute the rank alignment metric between conditional feature importance and ranked feature importance.

Parameters

conditional_feature_importanceConditionalImportances

The conditional feature importance values.

ranked_feature_importanceImportances

The ranked feature importance values.

aggregationbool, optional

If True, aggregate the results, by default True.

Returns

float

The computed rank alignment metric.