R/calculate_convergence.R

In missForestPredict: Missing Value Imputation using Random Forest for Prediction Settings

#' Calculates convergence based on NMSE
#'
#' Calculates convergence based on NMSE. Details on the convergence criterion calculation are
#' provided in the package vignettes.
#'
#' @param err dataframe containing OOB or apparent errors for each iteration.
#' @param weights vector of weights in the same format as for the \code{missForest} function.
#'
#' @return A list with elements
#'     \item{\code{converged}}{boolean indicating if the algorithm has converged (TRUE) or not (FALSE)}
#'     \item{\code{measure_old}}{the total error of the previous iteration}
#'     \item{\code{measure_new}}{the total error of the last iteration}
#' @export

calculate_convergence <- function(err, weights){

  iter <- max(err$iteration[!is.na(err$NMSE)])

  NMSE_err_new <- weighted.mean(err[err$iteration == iter,"NMSE"],
                                w = weights)
  NMSE_err_old <- weighted.mean(err[err$iteration == iter - 1,"NMSE"],
                                w = weights)

  converged <- NMSE_err_new >= NMSE_err_old

  return(list(converged = converged,
              measure_old = NMSE_err_old,
              measure_new = NMSE_err_new))

}