TrimMeanLeveneTest: Test for equality of variance based on trimmed-mean based...

In diffMeanVar: Detecting Gene Probes with Different Means or Variances Between Two Groups

Description

Test for equality of variance based on trimmed-mean based Levene test.

Usage

TrimMeanLeveneTest(value, group, trim.alpha = 0.25)

Arguments

value	numeric. Measurements to be compared between two groups.
group	numeric. Subject's group membership. Must be binary (i.e., taking values 0 or 1).
trim.alpha	numeric. Indicating proportion of data points to be trimmed from both ends before calculating sample mean.

Value

A list with 2 elements:

stat	test statistic value
pval	pvalue of the score test

Author(s)

Xuan Li <[e-mail]>, Weiliang Qiu <[e-mail]>, Yuejiao Fu <[e-mail]>, Xiaogang Wang <[e-mail]>

References

Brown MB and Forsythe AB (1974) Robust Tests for Equality of Variances. Journal of the American Statistical Association, 69, 364-367.

Examples

    # generate simulated data set from t distribution
    set.seed(1234567)
    es.sim = genSimData.tDistr(nCpGs = 100, nCases = 20, nControls = 20,
      df0 = 10, ncp0 = 0, df1 = 6, ncp1 = 2.393, testPara = "var",
      eps = 1.0e-3, applier = lapply) 
    print(es.sim)
    print(exprs(es.sim)[1:2,1:3])

    # do AW score test for the first probe
    dat = exprs(es.sim)
    pDat = pData(es.sim)
    print(pDat[1:2,])

    res = TrimMeanLeveneTest(value = dat[1,], group = pDat$memSubj)
    print(names(res))
    print(res)

diffMeanVar documentation built on May 2, 2019, 2:54 a.m.