R/calcFx.R

In PoweR: Computation of Power and Level Tables for Hypothesis Tests

calcFx <- function(pval.mat, x = c(seq(0.001, 0.009, by = 0.001), seq(0.010, 0.985, by = 0.005), seq(0.990, 0.999, by = 0.001))) {
# this function allows us to compute Fxi from the matrix of p-values obtained by many.pval()

  ## retrieve informations from matrix.pval$pvals computed by many.pval()
    nbstats <- ncol(pval.mat)
    N <- nrow(pval.mat)
    law <- unique(rownames(pval.mat))
    if (length(law) > 1) stop("The row names of 'pval.mat' should all be the same!")
    statnames <- colnames(pval.mat)
    
    Fx.mat <- matrix(0.0, nrow = length(x), ncol = nbstats) 
    for (i in 1:nbstats) {
        Fx.mat[,i] <- (.C("calcfx", as.double(pval.mat[, i]), as.integer(N), as.double(x),
                          as.integer(length(x)), fx = as.double(rep(0.0, length(x))), PACKAGE = "PoweR"))$fx
    }
    
    return(structure(list(Fx.mat = Fx.mat, x = x, law = law, statnames = statnames, N = N), class = c("Fx", "pvalue", "discrepancy", "sizepower")))
}