PKMeans {ORKM}

Power K-means clustering algorithm for single view data

Description

The power K-means algorithm is a generalization of the Lloyd algorithm, which approximates the ordinary K-means algorithm by a majorization-minimization method with the descent properties and lower complexity of the Lloyd algorithm. The power K-means embeds the K-means problem into a series of better performing problems. These smooth intermediate problems have a smoother objective function and tend to guide the clustering to find a global minimum with the K-means as the objective. The method has the same iteration complexity as Lloyd's algorithm, reduces sensitivity to initialization, and greatly improves algorithm performance in the high-dimensional case.

Usage

PKMeans(X, K, yitapower, sm, max.m, truere, method = 0)

Arguments

X

is the data matrix

K

is the number of cluster

yitapower

is the regularized parameter

sm

is the banlance parameter

max.m

is the max iter

truere

is the ture label in data set

method

is the caluate the NMI

Value

center,NMI,result

Author(s)

Miao Yu

Examples

library(MASS)   
  yitapower=0.5;K=3;sm=0.5;max.m=100;n1=n2=n3=70
  X1<-rnorm(n1,20,2);X2<-rnorm(n2,25,1.5);X3<-rnorm(n3,30,2) 
  Xv<-c(X1,X2,X3)
  data<-matrix(Xv,n1+n2+n3,2)
  data[1:70,2]<-1;data[71:140,2]<-2;data[141:210,2]<-3
  truere=data[,2]
  X11<-matrix(data[,1],n1+n2+n3,1) 
  PKMeans(X=X11,K=K,yitapower=yitapower,sm=sm,max.m=max.m,truere=truere,method=0)