Saved in:
Bibliographic Details
Main Authors: Paillard, Joseph, Lobo, Angel Reyero, Kolodyazhniy, Vitaliy, Thirion, Bertrand, Engemann, Denis A.
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2408.13002
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910243510812672
author Paillard, Joseph
Lobo, Angel Reyero
Kolodyazhniy, Vitaliy
Thirion, Bertrand
Engemann, Denis A.
author_facet Paillard, Joseph
Lobo, Angel Reyero
Kolodyazhniy, Vitaliy
Thirion, Bertrand
Engemann, Denis A.
contents Causal machine learning holds promise for estimating individual treatment effects from complex data. For successful real-world applications of machine learning methods, it is of paramount importance to obtain reliable insights into which variables drive heterogeneity in the response to treatment. We propose PermuCATE, an algorithm based on the Conditional Permutation Importance (CPI) method, for statistically rigorous global variable importance assessment in the estimation of the Conditional Average Treatment Effect (CATE). Theoretical analysis of the finite sample regime and empirical studies show that PermuCATE has lower variance than the Leave-One-Covariate-Out (LOCO) reference method and provides a reliable measure of variable importance. This property increases statistical power, which is crucial for causal inference in the limited-data regime common to biomedical applications. We empirically demonstrate the benefits of PermuCATE in simulated and real-world health datasets, including settings with up to hundreds of correlated variables.
format Preprint
id arxiv_https___arxiv_org_abs_2408_13002
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Measuring Variable Importance in Heterogeneous Treatment Effects with Confidence
Paillard, Joseph
Lobo, Angel Reyero
Kolodyazhniy, Vitaliy
Thirion, Bertrand
Engemann, Denis A.
Machine Learning
Causal machine learning holds promise for estimating individual treatment effects from complex data. For successful real-world applications of machine learning methods, it is of paramount importance to obtain reliable insights into which variables drive heterogeneity in the response to treatment. We propose PermuCATE, an algorithm based on the Conditional Permutation Importance (CPI) method, for statistically rigorous global variable importance assessment in the estimation of the Conditional Average Treatment Effect (CATE). Theoretical analysis of the finite sample regime and empirical studies show that PermuCATE has lower variance than the Leave-One-Covariate-Out (LOCO) reference method and provides a reliable measure of variable importance. This property increases statistical power, which is crucial for causal inference in the limited-data regime common to biomedical applications. We empirically demonstrate the benefits of PermuCATE in simulated and real-world health datasets, including settings with up to hundreds of correlated variables.
title Measuring Variable Importance in Heterogeneous Treatment Effects with Confidence
topic Machine Learning
url https://arxiv.org/abs/2408.13002