R/simPG2.R

In ipsfs: Intuitionistic, Pythagorean, and Spherical Fuzzy Similarity Measure

#'@title PFS similarity measure simPG2
#'@description PFS similarity measure values using simPG2 computation technique with membership, and non-membership values of two objects or set of objects.
#'@param ma PFS membership values for the data set x computed using either triangular or trapezoidal or guassian membership function
#'@param na PFS non-membership values for the data set x computed using either Sugeno and Terano's  or Yager's non-membership function
#'@param mb PFS membership values for the data set y computed using either triangular or trapezoidal or guassian membership function
#'@param nb PFS non-membership values for the data set y computed using either Sugeno and Terano's  or Yager's non-membership function
#'@param p Lp norm values for measuring the p-norm distance between x and y, values range from 1 to 5
#'@param l Level of uncertainty values, values range from 1 to 10
#'@param t Level of uncertainty values, values range from 1 to 10
#'@param k A constant value, considered as 1
#'@return The PFS similarity values of data set y with data set x
#'@references X. Peng and H. Garg. Multiparametric similarity measures on pythagorean fuzzy sets with applications to pattern recognition. Applied Intelligence, 49(12):4058 - 4096, 2019.
#'@examples
#'x<-matrix(c(12,9,14,11,21,16,15,24,20,17,14,11),nrow=4)
#'y<-matrix(c(11,21,6),nrow=1)
#'a<-mn(x)
#'b<-std(x)
#'a1<-mn(y)
#'b1<-std(y)
#'lam<-0.5
#'ma<-memG(a,b,x)
#'na<-nonmemS(ma,lam)
#'mb<-memG(a1,b1,y)
#'nb<-nonmemS(mb,lam)
#'p<-2
#'l<-2
#'t<-2
#'k<-1
#'simPG2(ma,na,mb,nb,p,l,t,k)
#'#[1] 0.5203669 0.5000073 0.7998594 0.7998594
#'@export
simPG2<-function(ma,na,mb,nb,p,l,t,k){
  c<-matrix(0,nrow=nrow(ma),ncol=ncol(ma))
  for (i in 1:nrow(ma)) {
    for(j in 1:ncol(ma))
      c[i,j]<-max((abs(((l-1)*(ma[i,j]^2-mb[k,j]^2))-(na[i,j]^2-nb[k,j]^2))^p),(abs(((l-t)*(na[i,j]^2-nb[k,j]^2))-(t*(ma[i,j]^2-mb[k,j]^2)))^p))
  }
  for(j in 1:ncol(c)){
    sum<-1-((rowSums(c)/(ncol(c)*((l)^p)))^(1/p))
  }
  sum
}