SHAP Values

SHapley Additive exPlanations. A game theory-based approach that assigns each feature an importance value for a particular prediction, providing consistent and locally accurate explanations.

A way to show which factors mattered most in an AI's decision. For example, SHAP might show that a loan was denied 40% because of income, 30% because of debt and 30% because of credit history.

From a governance perspective, SHAP values provide auditable evidence of how AI models make decisions. Incorporating SHAP analysis into your model documentation process strengthens your ability to demonstrate transparency and fairness.