Abstract:In this paper, we study distance covariance, Hilbert–Schmidt covariance (aka Hilbert–Schmidt independence criterion [In Advances in Neural Information Processing Systems (2008) 585–592]) and related independence tests under the high dimensional scenario. We show that the sample distance/Hilbert–Schmidt covariance between two random vectors can be approximated by the sum of squared componentwise sample cross-covariances up to an asymptotically constant factor, which indicates that the standard distance/Hilbert–Schmidt covariance based test can only capture linear dependence in high dimension. Under the assumption that the components within each high dimensional vector are weakly dependent, the distance correlation based $t$ test developed by Székely and Rizzo (J. Multivariate Anal. 117 (2013) 193–213) for independence is shown to have trivial limiting power when the two random vectors are nonlinearly dependent but component-wisely uncorrelated. This new and surprising phenomenon, which seems to be discovered and carefully studied for the first time, is further confirmed in our simulation study. As a remedy, we propose tests based on an aggregation of marginal sample distance/Hilbert–Schmidt covariances and show their superior power behavior against their joint counterparts in simulations. We further extend the distance correlation based $t$ test to those based on Hilbert–Schmidt covariance and marginal distance/Hilbert–Schmidt covariance. A novel unified approach is developed to analyze the studentized sample distance/Hilbert–Schmidt covariance as well as the studentized sample marginal distance covariance under both null and alternative hypothesis. Our theoretical and simulation results shed light on the limitation of distance/Hilbert–Schmidt covariance when used jointly in the high dimensional setting and suggest the aggregation of marginal distance/Hilbert–Schmidt covariance as a useful alternative.

Multivariate differential association analysis

Equivalence of Kernel Machine Regression and Kernel Distance Covariance for Multidimensional Trait Association Studies

Generalized kernel two-sample tests

A nonparametric Bayesian test of dependence

dCCA: detecting differential covariation patterns between two types of high-throughput omics data

Covariate-Adjusted Inference for Differential Analysis of High-Dimensional Networks

CONDITIONAL DISTANCE CORRELATION.

Set-based differential covariance testing for high-throughput data

A High-Dimensional Test for Multivariate Analysis of Variance under a Low-Dimensional Factor Structure

Two-Sample Test of High Dimensional Means under Dependence

A Kernel-Based Measure for Conditional Mean Dependence

A low variance consistent test of relative dependency

Two-Sample Test for Sparse High Dimensional Multinomial Distributions

Distance-based and RKHS-based dependence metrics in high dimension

High-dimensional sparse MANOVA

A machine learning-based approach for estimating and testing associations with multivariate outcomes

Multiple multi-sample testing under arbitrary covariance dependency

Set-based Differential Covariance Testing for Genomics.

On the Test of Covariance Between Two High-Dimensional Random Vectors

Multiple two-sample testing under arbitrary covariance dependency with an application in imaging mass spectrometry

A distance based two-sample test of means difference for multivariate datasets