Wasserstein Conditional Independence Testing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Warren, Andrew
Format: Preprint
Published: 2021
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929231185838080
author Warren, Andrew
author_facet Warren, Andrew
contents We introduce a test for the conditional independence of random variables $X$ and $Y$ given a random variable $Z$, specifically by sampling from the joint distribution $(X,Y,Z)$, binning the support of the distribution of $Z$, and conducting multiple $p$-Wasserstein two-sample tests. Under a $p$-Wasserstein Lipschitz assumption on the conditional distributions $\mathcal{L}_{X|Z}$, $\mathcal{L}_{Y|Z}$, and $\mathcal{L}_{(X,Y)|Z}$, we show that it is possible to control the Type I and Type II error of this test, and give examples of explicit finite-sample error bounds in the case where the distribution of $Z$ has compact support.
format Preprint
id arxiv_https___arxiv_org_abs_2107_14184
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Wasserstein Conditional Independence Testing
Warren, Andrew
Statistics Theory
Optimization and Control
62G10 (Primary), 49Q22 (Secondary)
We introduce a test for the conditional independence of random variables $X$ and $Y$ given a random variable $Z$, specifically by sampling from the joint distribution $(X,Y,Z)$, binning the support of the distribution of $Z$, and conducting multiple $p$-Wasserstein two-sample tests. Under a $p$-Wasserstein Lipschitz assumption on the conditional distributions $\mathcal{L}_{X|Z}$, $\mathcal{L}_{Y|Z}$, and $\mathcal{L}_{(X,Y)|Z}$, we show that it is possible to control the Type I and Type II error of this test, and give examples of explicit finite-sample error bounds in the case where the distribution of $Z$ has compact support.
title Wasserstein Conditional Independence Testing
topic Statistics Theory
Optimization and Control
62G10 (Primary), 49Q22 (Secondary)
url https://arxiv.org/abs/2107.14184