Explain Model Action Set

The shapleyExplainer action calculates Shapley values for predictive models.

Shapley values are a measure of variable importance for a single observation. They are an additive measure of importance whose sum equals the predicted output from the model that you are explaining.

The shapleyExplainer action uses a method called hyperparameterized Shapley value estimation (HyperSHAP) to calculate the Shapley values.