The conditional permutation test for independence while controlling for confounders

DOI10.1111/RSSB.12340MaRDI QIDQ5117627zbMATH OpenOpenAlexWikidataFDO

Authors Thomas B. Berrett, Yi Wang, Rina Foygel Barber, Richard Samworth

Publication date 26 August 2020

Published in Journal of the Royal Statistical Society Series B: Statistical Methodology (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1807.05405

zbMATH Keywords

conditional independence testingexchangeabilitymodel misspecificationmodel-X knockoffspermutation testssemisupervised learning

Mathematics Subject Classification ID

Foundations and philosophical topics in statistics (62A01) Measures of association (correlation, canonical correlation, etc.) (62H20)

Abstract: We propose a general new method, the conditional permutation test, for testing the conditional independence of variables $X$ and $Y$ given a potentially high-dimensional random vector $Z$ that may contain confounding factors. The proposed test permutes entries of $X$ non-uniformly, so as to respect the existing dependence between $X$ and $Z$ and thus account for the presence of these confounders. Like the conditional randomization test of Cand`es et al. (2018), our test relies on the availability of an approximation to the distribution of $XmidZ$. While Cand`es et al. (2018)'s test uses this estimate to draw new $X$ values, for our test we use this approximation to design an appropriate non-uniform distribution on permutations of the $X$ values already seen in the true data. We provide an efficient Markov Chain Monte Carlo sampler for the implementation of our method, and establish bounds on the Type I error in terms of the error in the approximation of the conditional distribution of $XmidZ$, finding that, for the worst case test statistic, the inflation in Type I error of the conditional permutation test is no larger than that of the conditional randomization test. We validate these theoretical results with experiments on simulated data and on the Capital Bikeshare data set.

Recommendations

• Minimax optimal conditional independence testing
• The Hardness of Conditional Independence Testing and the Generalised Covariance Measure
• Test for conditional independence with application to conditional screening
• Strongly consistent nonparametric tests of conditional independence
• Testing conditional independence via integrating-up transform

Cited in

(45)

• Testing conditional independence in supervised learning algorithms
• A simple measure of conditional dependence
• Optimal rates for independence testing via U-statistic permutation tests
• Asymptotic distributions of high-dimensional distance correlation inference
• Minimax optimal conditional independence testing
• Unrestricted permutation forces extrapolation: variable importance requires at least one more model, or there is no free variable importance
• Testing goodness-of-fit and conditional independence with approximate co-sufficient sampling
• On the power of conditional independence testing under model-X
• Local permutation tests for conditional independence
• InfoGram and admissible machine learning
• Relaxing the assumptions of knockoffs by conditioning
• scientific article; zbMATH DE number 7626800 (Why is no real title available?)
• The Holdout Randomization Test for Feature Selection in Black Box Models
• Hypothesis testing in adaptively sampled data: ART to maximize power beyond \textit{iid }sampling
• Detecting distributional differences in labeled sequence data with application to tropical cyclone satellite imagery
• Analysis of conditional randomisation and permutation schemes with application to conditional independence testing
• Learning to increase the power of conditional randomization tests
• Nonlinear directed acyclic graph estimation based on the kernel partial correlation coefficient
• What is a Randomization Test?
• On Azadkia-Chatterjee's conditional dependence coefficient
• Anytime-Valid Tests of Conditional Independence Under Model-X
• A double-robust test for high-dimensional gene coexpression networks conditioning on clinical information
• Reconciling model-X and doubly robust approaches to conditional independence testing
• A zero-estimator approach for estimating the signal level in a high-dimensional model-free setting
• Studentized permutation method for comparing two restricted mean survival times with small sample from randomized trials
• A survey of some recent developments in measures of association
• Permuting density-based empirical likelihood ratio test statistics
• Markov chain Monte Carlo significance tests
• Randomization tests for conditional group symmetry
• Double robust conditional independence test for novel biomarkers given established risk factors with survival data
• Sequential Monte Carlo testing by betting
• Rank-transformed subsampling: inference for multiple data splitting and exchangeable p-values
• Estimating causal effects under non-individualistic treatments due to network entanglement
• Covariate-adaptive randomization inference in matched designs
• Pseudo-likelihood ratio screening based on network data with applications
• Smooth and Probabilistic PARAFAC Model with Auxiliary Covariates
• Model-Based Causal Feature Selection for General Response Types
• Residual permutation test for regression coefficient testing
• Kernel copula density estimation of Hellinger correlation
• Conditional independence testing for discrete distributions: beyond ^2- and G-tests
• The projected covariance measure for assumption-lean variable significance testing
• Association and independence test for random objects
• Approximate co-sufficient sampling with regularization
• An adaptive-to-model distributed hybrid test for conditional independence
• Nonparametric methods controlling the median of the false discovery proportion

This page was built for publication: The conditional permutation test for independence while controlling for confounders

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5117627)