SHAP methods (SHapley Additive exPlanations)

Global Explainability

Level of interpretation that aggregates local explanations over a dataset to understand the general behavior of the model and the average importance of features.

← Wstecz