Limiting distributions of likelihood ratio test for independence of components for high-dimensional normal vectors

Yongcheng Qi, Fang Wang, Lin Zhang

Research output: Contribution to journalArticle

Abstract

Consider a p-variate normal random vector. We are interested in the limiting distributions of likelihood ratio test (LRT) statistics for testing the independence of its grouped components based on a random sample of size n. In classical multivariate analysis, the dimension p is fixed or relatively small, and the limiting distribution of the LRT is a chi-square distribution. When p goes to infinity, the chi-square approximation to the classical LRT statistic may be invalid. In this paper, we prove that the LRT statistic converges to a normal distribution under quite general conditions when p goes to infinity. We propose an adjusted test statistic which has a chi-square limit in general. Our comparison study indicates that the adjusted test statistic outperforms among the three approximations in terms of sizes. We also report some numerical results to compare the performance of our approaches and other methods in the literature.

Original languageEnglish (US)
Pages (from-to)911-946
Number of pages36
JournalAnnals of the Institute of Statistical Mathematics
Volume71
Issue number4
DOIs
StatePublished - Aug 7 2019

Fingerprint

Likelihood Ratio Test Statistic
Normal vector
Likelihood Ratio Test
Limiting Distribution
High-dimensional
Chi-square
Test Statistic
Infinity
Chi-square Distribution
Multivariate Analysis
Approximation
Random Vector
Gaussian distribution
Converge
Numerical Results
Testing
Independence

Keywords

  • Central limit theorem
  • Chi-square approximation
  • Covariance matrix
  • High-dimensional normal vector
  • Independence
  • Likelihood ratio test

Cite this

Limiting distributions of likelihood ratio test for independence of components for high-dimensional normal vectors. / Qi, Yongcheng; Wang, Fang; Zhang, Lin.

In: Annals of the Institute of Statistical Mathematics, Vol. 71, No. 4, 07.08.2019, p. 911-946.

Research output: Contribution to journalArticle

@article{584ec384e54f4ad68c8cbb2e826a7728,
title = "Limiting distributions of likelihood ratio test for independence of components for high-dimensional normal vectors",
abstract = "Consider a p-variate normal random vector. We are interested in the limiting distributions of likelihood ratio test (LRT) statistics for testing the independence of its grouped components based on a random sample of size n. In classical multivariate analysis, the dimension p is fixed or relatively small, and the limiting distribution of the LRT is a chi-square distribution. When p goes to infinity, the chi-square approximation to the classical LRT statistic may be invalid. In this paper, we prove that the LRT statistic converges to a normal distribution under quite general conditions when p goes to infinity. We propose an adjusted test statistic which has a chi-square limit in general. Our comparison study indicates that the adjusted test statistic outperforms among the three approximations in terms of sizes. We also report some numerical results to compare the performance of our approaches and other methods in the literature.",
keywords = "Central limit theorem, Chi-square approximation, Covariance matrix, High-dimensional normal vector, Independence, Likelihood ratio test",
author = "Yongcheng Qi and Fang Wang and Lin Zhang",
year = "2019",
month = "8",
day = "7",
doi = "10.1007/s10463-018-0666-9",
language = "English (US)",
volume = "71",
pages = "911--946",
journal = "Annals of the Institute of Statistical Mathematics",
issn = "0020-3157",
publisher = "Springer Netherlands",
number = "4",

}

TY - JOUR

T1 - Limiting distributions of likelihood ratio test for independence of components for high-dimensional normal vectors

AU - Qi, Yongcheng

AU - Wang, Fang

AU - Zhang, Lin

PY - 2019/8/7

Y1 - 2019/8/7

N2 - Consider a p-variate normal random vector. We are interested in the limiting distributions of likelihood ratio test (LRT) statistics for testing the independence of its grouped components based on a random sample of size n. In classical multivariate analysis, the dimension p is fixed or relatively small, and the limiting distribution of the LRT is a chi-square distribution. When p goes to infinity, the chi-square approximation to the classical LRT statistic may be invalid. In this paper, we prove that the LRT statistic converges to a normal distribution under quite general conditions when p goes to infinity. We propose an adjusted test statistic which has a chi-square limit in general. Our comparison study indicates that the adjusted test statistic outperforms among the three approximations in terms of sizes. We also report some numerical results to compare the performance of our approaches and other methods in the literature.

AB - Consider a p-variate normal random vector. We are interested in the limiting distributions of likelihood ratio test (LRT) statistics for testing the independence of its grouped components based on a random sample of size n. In classical multivariate analysis, the dimension p is fixed or relatively small, and the limiting distribution of the LRT is a chi-square distribution. When p goes to infinity, the chi-square approximation to the classical LRT statistic may be invalid. In this paper, we prove that the LRT statistic converges to a normal distribution under quite general conditions when p goes to infinity. We propose an adjusted test statistic which has a chi-square limit in general. Our comparison study indicates that the adjusted test statistic outperforms among the three approximations in terms of sizes. We also report some numerical results to compare the performance of our approaches and other methods in the literature.

KW - Central limit theorem

KW - Chi-square approximation

KW - Covariance matrix

KW - High-dimensional normal vector

KW - Independence

KW - Likelihood ratio test

UR - http://www.scopus.com/inward/record.url?scp=85046891128&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046891128&partnerID=8YFLogxK

U2 - 10.1007/s10463-018-0666-9

DO - 10.1007/s10463-018-0666-9

M3 - Article

VL - 71

SP - 911

EP - 946

JO - Annals of the Institute of Statistical Mathematics

JF - Annals of the Institute of Statistical Mathematics

SN - 0020-3157

IS - 4

ER -