Post-hoc explainability

Decision Tree Surrogates

Technique that trains an interpretable decision tree to approximate the behavior of a complex model by using the model's predictions as targets.

← Wstecz