Premium
An accurate test for the equality of covariance matrices from decomposable graphical Gaussian models
Author(s) -
Wu Yanyan,
Massam Hélène,
Wong Augustine
Publication year - 2014
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.1002/cjs.11201
Subject(s) - mathematics , covariance , gaussian , covariance matrix , statistic , statistics , estimation of covariance matrices , graphical model , markov chain , covariance function , test statistic , statistical hypothesis testing , physics , quantum mechanics
This paper derives a saddlepoint based approximation for the cumulative distribution function of the Bartlett–Box M‐statistic that tests the equality of covariance matrices for several samples from graphical Gaussian models Markov with respect to a decomposable graph G . The proposed saddlepoint‐based method has third‐order accuracy ( O ( n − 3 / 2 ) ). Simulation results show that the proposed method has extremely good coverage properties even when the sample size is small. We apply our method to the well‐known Call Centre data set and show that the covariance matrix is not constant through time. The Canadian Journal of Statistics 42: 61–77; 2014 © 2014 Statistical Society of Canada