SHapley Additive exPlanations (SHAP): A unified measure of feature importance for any given prediction, it can be used to create human-understandable explanations of the model’s decisions.