A General Framework for Inference on Algorithm-Agnostic Variable Importance

Brian D. Williamson

doi:10.1080/01621459.2021.2003200

Return

A General Framework for Inference on Algorithm-Agnostic Variable Importance

2022-01-05

Brian D. Williamson *

Peter B. Gilbert

Noah Simon

Marco Carone

DOI：10.1080/01621459.2021.2003200

Original

Original request for help

Save

Abstract

En 中文

In many applications, it is of interest to assess the relative contribution of features (or subsets of features) toward the goal of predicting a response-in other words, to gauge the variable importance of features. Most recent work on variable importance assessment has focused on describing the importance of features within the confines of a given prediction algorithm. However, such assessment does not necessarily characterize the prediction potential of features, and may provide a misleading reflection of the intrinsic value of these features. To address this limitation, we propose a general framework for nonparametric inference on interpretable algorithm-agnostic variable importance. We define variable importance as a population-level contrast between the oracle predictiveness of all available features versus all features except those under consideration. We propose a nonparametric efficient estimation procedure that allows the construction of valid confidence intervals, even when machine learning techniques are used. We also outline a valid strategy for testing the null importance hypothesis. Through simulations, we show that our proposal has good operating characteristics, and we illustrate its use with data from a study of an antibody against HIV-1 infection. Supplementary materials for this article are available online.

Keywords:

Machine learning

Statistical inference

Targeted learning

Variable importance