equalCovs {equalCovs}

Testing the equality of two covariance matrices.

Description

R code for testing the equality of two covariance matrices, used in paper "Two sample tests for high dimensional covariance matrices".

Usage

equalCovs(sam1, sam2, size1, size2)

Arguments

sam1

First sample, it must be array with structure size1*p, p is the dimension of data.

sam2

Second sample, it must be array with structure size2*p, p is the dimension of data.

size1

sample size of first sample

size2

sample size of second sample

Value

test statistics and p-values

test_stat

test statistics

pvalue

p-values

Author(s)

Jun Li and Song Xi Chen

Examples

library(mvtnorm)
p<-700 # the dimension of multivariate

theta1<-2
theta2<-1
mat1<-diag(theta1,p-1)
mat2<-diag(theta1+theta1*theta2,p-1)
mat3<-diag(theta2,p-2)

mat1<-rbind(mat1,rep(0,p-1))
mat2<-rbind(mat2,rep(0,p-1))
mat3<-rbind(mat3,rep(0,p-2),rep(0,p-2))

mat1<-cbind(rep(0,p),mat1)
mat2<-cbind(rep(0,p),mat2)
mat3<-cbind(rep(0,p),rep(0,p),mat3)
sigma1<-mat1+t(mat1)+diag(1+theta1^2,p)
sigma2<-mat2+t(mat2)+mat3+t(mat3)+diag(1+theta1^2+theta2^2,p)

size1<-80
size2<-80
sam1<-rmvnorm(size1,runif(p,0,5),sigma1) # generate the samples
sam2<-rmvnorm(size2,runif(p,-3,3),sigma2)

equalCovs(sam1,sam2,size1,size2)