R/function-defuzzify.R

In FuzzyAHP: (Fuzzy) AHP Calculation

#' Function to defuzzify fuzzy data
#'
#' @description
#' This function defuzzifies \code{\linkS4class{FuzzyData}} into single value.
#' The \code{\linkS4class{FuzzyData}} can only contain one fuzzy number, otherwise the defuzzification is not possible.
#'
#' @details
#' The triangular fuzzy numbers \eqn{\tilde{A}}{A} represented as triplet \eqn{[a1,a2,a3]} are defuzzified using these
#' methods:
#' \code{"Yager"}:
#' \deqn{def (\tilde{A}) = \frac{(a2-a1)(a1+2/3(a2-a1)+(a3-a2)(a2+1/3(a3-a2))}{(a2-a1)+(a3-a2))}}{def(A) = ((a2-a1)(a1+2/3(a2-a1)+(a3-a2)(a2+1/3(a3-a2))) / ((a2-a1)+(a3-a2)))}
#' \code{"modalvalue"}:
#' \deqn{def (\tilde{A}) = a2}{def(A) = a2}
#' \code{"modalValueDominancy"}:
#' \deqn{def(\tilde{A}) = (a1 + 4*a2 + a3)/6}{def(A) = (a1 + 4*a2 + a3)/6}
#' \code{"mean"}:
#' \deqn{def(\tilde{A}) = (a1 + a2 + a3)/3}{def(A) = (a1 + a2 + a3)/3}.
#'
#' @param fuzzyData A \code{\linkS4class{FuzzyData}}
#' @param type A \code{"character"} representing type of defuzzification. Currently implemented methods are
#' \code{"Yager"}, \code{"modalValue"}, \code{"modalValueDominancy"}, \code{"mean"}. The default value is \code{"mean"}.
#'
#' @return A numeric value of defuzzified value, based on deffuzification method.
#'
#' @export
#' @rdname defuzziffy-methods
#' @name defuzziffy
setGeneric("defuzziffy",
           function(fuzzyData, type = "mean") standardGeneric("defuzziffy"))

#' @rdname defuzziffy-methods
#' @aliases defuzziffy,FuzzyData,character-method
setMethod(
  f="defuzziffy",
  signature(fuzzyData = "FuzzyData"),
  definition=function(fuzzyData, type)
  {
    if(typeof(type) != "character"){
      stop("Variable type must be character!")
    }

    if(ncol(fuzzyData@fnMin)>1){
      stop("Defuzzification is only possible for datasets representing one fuzzy number.")
    }

    numberRows = nrow(fuzzyData@fnModal)

    result = matrix(NA, nrow = numberRows, ncol = 1)

    if(type == "Yager"){
      result[,1] = (((fuzzyData@fnModal[,1]-fuzzyData@fnMin[,1])*(fuzzyData@fnMin[,1]+(2/3)*(fuzzyData@fnModal[,1]-fuzzyData@fnMin[,1]))+((fuzzyData@fnMax[,1]-fuzzyData@fnModal[,1])*(fuzzyData@fnModal[,1]+(1/3)*(fuzzyData@fnMax[,1]-fuzzyData@fnModal[,1]))))/((fuzzyData@fnModal[,1]-fuzzyData@fnMin[,1])+(fuzzyData@fnMax[,1]-fuzzyData@fnModal[,1])))

    }else if(type == "modalValue"){
      result[,1] = fuzzyData@fnModal[,1]

    }else if(type == "modalValueDominancy"){
      result[,1] = (fuzzyData@fnMin[,1] + 4*fuzzyData@fnModal[,1] + fuzzyData@fnMax[,1])/6

    }else if (type == "mean"){
      result[,1] = (fuzzyData@fnMin[,1] + fuzzyData@fnModal[,1] + fuzzyData@fnMax[,1])/3

    }else{
      warning("Name of deffuzification method not recognized. Using default method mean of triganular fuzzy number instead.")
      result[,1] = (fuzzyData@fnMin[,1] + fuzzyData@fnModal[,1] + fuzzyData@fnMax[,1])/3
    }

    return(result)
  }
)