R/DIC.R

In sbfnk/RBi.helpers: 'rbi' Helper Functions

#' @export
DIC <- function(x, ...) UseMethod("DIC") # nolint
##' @name DIC
##' @rdname DIC
##' @title Compute Deviance Information Criterion (DIC) for a libbi model
##'
##' @description
##' Computes the DIC of a libbi object containing Monte-Carlo samples. The
##'   effective number of parameters is calculated following Gelman et al.,
##'   Bayesian Data Analysis: Second Edition, 2004, p. 182.
##'
##' @param x a \code{libbi} object
##' @param bootstrap number of bootstrap samples to take, 0 to just take data
##' @param ... any parameters to be passed to `bi_read` (e.g., `burn`)
##' @return DIC
##' @importFrom stats var
##' @export
##' @examples
##' example_run <- rbi::bi_read(
##'   system.file(package = "rbi", "example_output.nc")
##' )
##' example_model_file <- system.file(package = "rbi", "PZ.bi")
##' example_bi <- rbi::attach_data(
##'   rbi::libbi(example_model_file), "output", example_run
##' )
##' DIC(example_bi)
##' @author Sebastian Funk
DIC.libbi <- function(x, bootstrap = 0, ...) {
  res <- bi_read(x, ...)

  ## sample mean deviance
  mean_d <- mean(-2 * res[["loglikelihood"]]$value)

  ## effective number of parameters
  pd <- stats::var(-2 * res[["loglikelihood"]]$value) / 2

  ## DIC
  return(mean_d + pd)
}