Description
Machine-learning predictions can be explained with SHAP values for model inspection and debugging. Data scientists can estimate feature contributions, create explanation plots, and compare model behavior on datasets. It processes model inputs that may contain sensitive training or user data.