SHAP Values

SHAP (SHapley Additive exPlanations) values provide local interpretability by assigning an importance value to each feature for a given prediction. They explain why a model gave a specific score for a specific input. SHAP is commonly used to justify AI decisions in regulated settings like credit approval or medical diagnoses.