R/evaluate_bootstrap.R

In cases: Stratified Evaluation of Subgroup Classification Accuracy

#' @importFrom boot boot
evaluate_bootstrap <- function(data,
                               contrast = define_contrast("raw"),
                               benchmark = 0.5,
                               alpha = 0.05,
                               alternative = "greater",
                               analysis = "co-primary",
                               transformation = "none",
                               regu = c(1, 1 / 2, 0),
                               pars = list(),
                               attrs = list()) {
  ## inference:
  stats <- data2stats(data, contrast = contrast, regu = regu)
  bst <- bootstrap_sample(data, contrast, regu, alternative, analysis, pars)
  bst_max <- apply(bst, 1, max)
  critval <- critval_bootstrap(alpha, alternative, bst_max)
  alpha_adj <- alpha_bootstrap(alpha, alternative, bst)

  ## output:
  stats2results(
    stats = stats,
    alpha = alpha,
    alpha_adj = alpha_adj,
    critval = critval,
    pval_fun = pval_bootstrap,
    pval_args = list(bst_max = bst_max),
    benchmark = benchmark,
    alternative = alternative,
    analysis = analysis,
    transformation = transformation,
    attrs = attrs
  )
}