Unrestricted permutation forces extrapolation: variable importance requires at least one more model, or there is no free variable importance

This paper reviews and advocates against the use of permute-and-predict (PaP) methods for interpreting black box functions. Methods such as the variable importance measures proposed for random forests, partial dependence plots, and individual conditional expectation plots remain popular because they...

Full description

Bibliographic Details

Published in:	Statistics and computing Vol. 31; no. 6
Main Authors:	Hooker, Giles, Mentch, Lucas, Zhou, Siyu
Format:	Journal Article
Language:	English
Published:	New York Springer US 01.11.2021 Springer Nature B.V
Subjects:	Artificial Intelligence Mathematics and Statistics Permutations Probability and Statistics in Computer Science Statistical Theory and Methods Statistics Statistics and Computing/Statistics Programs Feature importance Permutation Functional ANOVA Intelligibility ICE plot Interpretation Partial dependence
ISSN:	0960-3174, 1573-1375
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!