Knowledge MaxxingSHAP value
A SHAP value attributes a model's specific prediction for one example to each of its input features, showing how much each feature pushed the prediction up or down.
Prerequisite: Feature importance
What matters
Unlike overall feature importance, SHAP values explain one individual prediction, which is what lets you answer 'why did the model decide this for this specific case?'
Example
A SHAP explanation for one loan denial might show income contributed positively while a short credit history contributed negatively to that specific decision.
Common misconception
A SHAP value does not reveal the model's true internal reasoning process; it is a mathematical attribution method, and different explanation methods can disagree on the same prediction.
Recall question
How does a SHAP value differ from an overall feature-importance ranking?