R/trivRegion.R

In refreg: Conditional Multivariate Reference Regions

#' Trivariate reference region estimation
#'
#' This functions estimate a probabilistic/reference region for trivariate data.
#' It is based on a non parametric kernel density estimation. It can only be applied to
#' a trivRegr object, and for one single tau.
#'
#' @param fit A trivRegr object.
#' @param tau A number defining the desired coverage of the trivariate
#' reference region.
#' @return This function return a region containing a given percentage of trivariate data points.
#' @references Duong, T. (2019) ks: Kernel Smoothing. R package version 1.11.6. https://CRAN.R--project.org/package=ks.
#' @export
#' @importFrom "ks" "kde"

trivRegion <- function(fit, tau = 0.90) {
  kernel.3d <- kde(fit$trivres,
    H = diag(c(0.5, 0.5, 0.5)),
    approx.cont = TRUE, gridsize = c(49, 50, 51)
  )
  ks_hat <- predict(kernel.3d, x = as.matrix(fit$trivres))


  names(kernel.3d$cont) <- as.numeric(substr(names(kernel.3d$cont), 1, nchar(names(kernel.3d$cont)) - 1)) / 100
  k <- as.numeric(which(ks_hat < kernel.3d$cont[as.character(1 - tau)]))
  l <- list(
    kernel_fit = kernel.3d, predict_kernel = ks_hat, which_out = k,
    fit = fit, trivres = fit$trivres, k_limit = as.numeric(kernel.3d$cont[as.character(1 - tau)])
  )
  class(l) <- "trivRegion"
  return(l)
}