From local explanations to global understanding with explainable AI for trees

An explanation method for trees is presented that enables the computation of optimal local explanations for individual predictions, and the authors demonstrate their method on three medical datasets.