XploRe Help : xcfcgk

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library:	xclust
See also:	xfcme xchcme

Quantlet:		xcfcgk
Description:		performs a fuzzy Gustafson-Kessel cluster analysis.

Reference(s):

James C. Bezdek (1981). Pattern Recognition with Fuzzy Objective Function Algorithm. Plenum Press, New York. James C. Bezdek and Sankar K. Pal (1992). Fuzzy Models for Pattern Recognition. IEEE Press, New York.

Usage:	`{v,d,uu}=xcfcme(x,c,m,e)`
Input:
	`x`	n x p matrix of n row points to be clustered
	`c`	scalar representing the number of clusters
	`m`	scalar, determines the fuzziness of clustering (m>1)
	`e`	scalar, termination tolerance
Output:
	`v`	p x p matrix of cluster centers
	`d`	n x p matrix of distances
	`uu`	n x p matrix of result

Example:

proc(xx,yy1,yy2)=idxgraph(out,x) xx=matrix(max(x[,1])+1)*0 yy1=xx yy2=xx i=1 while(i <= max(x[,1])+1) xx[i]= i-1 yy1[i]= max(paf(out.uu'[,1], x[,1]==i-1)) yy2[i]= max(paf(out.uu'[,2], x[,1]==i-1)) i=i+1 endo endp library("xclust") library("plot") z=read("butterfly") x=z[,2:3] c=2 m=2 e=0.01 out=xcfcgk(x,c,m,e) //apply fuzzy gustafson-kessel clustering out.uu' tt=idxgraph(out,x) tyy1=tt.xx~tt.yy1 tyy2=tt.xx~tt.yy2 tyy1 = setmask(tyy1, "line", "red", "thin") tyy2 = setmask(tyy2, "line", "blue", "medium") disp = createdisplay(1,1) show(disp,1,1, tyy1, tyy2) strm = string("%1.0f", m) setgopt(disp,1,1, "title","Membership Function for Fuzzy GK with m="+ strm)

Result:

gives the partitions of membership functions and depicts
this membership function for m = 2.

Author: H. Sofyan, 20020114