Skip to content
ML Academy

Introduced on day 17

SHAP value

A Shapley-value-based attribution that splits the gap between one prediction and the baseline prediction across the features, additively.

In detail

SHAP values sum to the prediction minus the baseline and give local explanations; averages of absolute values give a global ranking. They describe the model: correlated features share credit in a method-dependent way, and a value is not the effect of changing the feature in the world (Lundberg and Lee, 2017; SHAP documentation).