htsq:

In shaileshtripathi/ssapbm: self-contained and competitive pathway-based methos (sub sample analysis of pathway-based methods.)

Usage

htsq(data, cls1, cls2, shrinkage = F)

Arguments

data
cls1
cls2
shrinkage

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (data, cls1, cls2, shrinkage = F) 
{
    meanS1 <- apply(data[, cls1], 1, mean)
    meanS2 <- apply(data[, cls2], 1, mean)
    covS1 <- covS2 <- NULL
    if ((nrow(data) - 1) > (ncol(data))) {
        shrinkage = T
    }
    if (shrinkage) {
        covS1 <- cov.shrink(t(data[, cls1]), verbose = F)
        covS2 <- cov.shrink(t(data[, cls2]), verbose = F)
    }
    else {
        covS1 <- cov(t(data[, cls1]))
        covS2 <- cov(t(data[, cls2]))
    }
    n1 = ncol(data[, cls1])
    n2 = ncol(data[, cls2])
    p <- nrow(data)
    covS = ((n1 - 1) * covS1 + (n2 - 1) * covS2)/(n1 + n2 - 2)
    covSinv <- solve(covS)
    htsqx <- (t(meanS1 - meanS2) %*% covSinv %*% (meanS1 - meanS2)) * 
        (n1 * n2/(n1 + n2))
    list(statistic = htsqx, df1 = p, df2 = n1 + n2 - p - 1, n1 = n1, 
        n2 = n2)
  }

shaileshtripathi/ssapbm documentation built on May 29, 2019, 8:06 p.m.