From unbiased MDI Feature Importance to Explainable AI for Trees
Machine Learning
2021-10-01 v4 Machine Learning
Computation
Abstract
We attempt to give a unifying view of the various recent attempts to (i) improve the interpretability of tree-based models and (ii) debias the the default variable-importance measure in random Forests, Gini importance. In particular, we demonstrate a common thread among the out-of-bag based bias correction methods and their connection to local explanation for trees. In addition, we point out a bias caused by the inclusion of inbag data in the newly developed explainable AI for trees algorithms.
Cite
@article{arxiv.2003.12043,
title = {From unbiased MDI Feature Importance to Explainable AI for Trees},
author = {Markus Loecher},
journal= {arXiv preprint arXiv:2003.12043},
year = {2021}
}
Comments
arXiv admin note: text overlap with arXiv:2003.02106