Chang.maxBoot.test: Two-sample covariance test (Chang et al. 2016)

In lingxuez/sLED: A Sparse Leading Eigenvalue Driven (sLED) Test for High-dimensional Matrices

Description

The two-sample test for high-dimensional covariance matrices in Chang, Zhou, Zhou, and Wang (2016) "Comparing Large Covariance Matrices under Weak Conditions on the Dependence Structure". Using a wild boostrap procedure, the test statistic is essentially the square root of the statistic proposed in Cai, Liu and Xia (2013).

Usage

Chang.maxBoot.test(X, Y, nresample = 1000, useMC = TRUE, mc.cores = 1)

Arguments

X	n1 by p matrix, observation of the first population, columns are features
Y	n2 by p matrix, observation of the second population, columns are features
nresample	the number of bootstraps to perform
useMC	logical variable indicating whether to use multicore parallelization. R packages parallel and doParallel are required if set to TRUE.
mc.cores	decide the number of cores to use when useMC is set to TRUE.

Value

A list with the following components:

test.stat	test statistic
test.stat.boot	bootstrap test statistics, a numeric vector with length nresample
pVal	bootstrap p-value

References

Chang, Zhou, Zhou, and Wang (2016) "Comparing Large Covariance Matrices under Weak Conditions on the Dependence Structure", arXiv preprint arXiv:1505.04493.

See Also

Cai.max.test(), LC.U.test(), WL.randProj.test(), Schott.Frob.test().

lingxuez/sLED documentation built on Oct. 6, 2017, 12:14 a.m.