R/cv_weight.R

In zhjx19/mathmodels: Comprehensive Mathematical Modeling Algorithms in R

#' Coefficient of Variation Weighting
#'
#' @description
#' Computes weights for indicators using the Coefficient of Variation (CV) method.
#' Weights are derived by normalizing the CV (standard deviation divided by mean)
#' for each indicator.
#'
#' @param data Numeric matrix or data frame with positive indicator data.
#'
#' @return Numeric vector of weights for the indicators, summing to 1.
#'
#' @details
#' The \code{cv_weight} function calculates weights using the CV method.
#' For each column in \code{data}, the CV is computed as the standard deviation
#' divided by the mean. Weights are obtained by normalizing the CVs to sum to 1.
#' This lightweight implementation uses base R and assumes all columns are numeric
#' indicators.
#'
#' @examples
#' X = data.frame(x1 = c(10, 20, 15), x2 = c(5, 10, 8))
#' cv_weight(X)
#'
#' @export
cv_weight = function(X) {
  # Compute weights using Coefficient of Variation
  # data: positive indicator data

  # Calculate mean and standard deviation
  means = apply(X, 2, mean, na.rm = TRUE)
  sds = apply(X, 2, sd, na.rm = TRUE)
  cv = sds / means     # Compute coefficient of variation
  # Normalize to weights
  cv / sum(cv)
}