nulldist.twosample: Two sample null distribution generator

In dnayyala/cramp: Covariance Matrix Testing using Random Projections

Two sample null distribution generator

Description

This function generates random samples and evaluates the test statistics for the two-sample covariance matrix test and reports a bootstrap random sample of average p-values. The generated values can be used to calculate the critical value for rejecting/accepting the null hypothesis

Usage

nulldist.twosample(
  n1,
  n2 = NULL,
  p,
  m = 1,
  nnull = 100,
  nrep = 100,
  mean1 = NULL,
  mean2 = NULL,
  parallel = TRUE,
  ncores = NULL
)

Arguments

n1	sample size of the data from the first group
n2	sample size of the data from the second group. If n2 = NULL (Default value), then n2 is set equal to n1
p	original dimension of your data
m	projected dimension of your data (Default value is 1)
nnull	number of repetitions of averages of pvalues
nrep	number of repetitions for each average of pvalues
mean1	(Optional) Mean vector of the first group for the null distribution
mean2	(Optional) Mean vector of the second group for the null distribution
parallel	Whether to parallelize or not (default is TRUE)
ncores	Number of cores to use (If NULL and parallel is TRUE, use 75% of all cores). If parallel = FALSE, this argument will be ignored.

Value

A list with three elements corresponding to different test statistics

Box-M

Box's M test statistic

Wald

Wald test statistic

Examples

## Not run: 
 nulldist.twosample(n = 20, p = 100, m = 5)

## End(Not run)

dnayyala/cramp documentation built on June 27, 2023, 1:34 p.m.