SHapley Additive exPlanations (SHAP)
  • is a game theoretic approach to explain the output of any machine learning model