R/simul.R

In nt-williams/simul: Fast Simultaneous Confidence Bands Based on the Efficient Influence Function and Multiplier Bootstrap

#' Find Simultaneous Confidence Band Critical Value
#'
#' @param x A list of parameter estimates.
#' @param eif A list of empirical efficient influence functions corresponding to the estimates in \code{x}.
#' @param nobs The number of observations.
#' @param reps The number of repetitions to use for the multiplier bootstrap, the default is 1e5.
#' @param level The confidence level for the critical value should be calculated for, the default is 0.95.
#'
#' @return The estimated critical value satisfying the requirements for a uniform confidence band around all estimates.
#' @export
#'
#' @examples
#' data(eif)
#' psi <- lapply(eif, function(x) mean(x))
#' n <- length(eif[[1]])
#' simul(psi, eif, n)
simul <- function(x, eif, nobs, reps = 1e5, level = 0.95) {
  multboot_checks(x, eif, nobs, level)
  mbs <- multBoot(x, eif, nobs, reps)
  return(quantile(mbs, level, names = FALSE))
}