kmeancluster.java

来自「it is Also Kmeans Algorithm in java.」· Java 代码 · 共 102 行

/*
 * To change this template, choose Tools | Templates
 * and open the template in the editor.
 */

package clusteranalysis;


public class KMeanCluster {

    private double[][] data;
    private int groups;
    private double[][] centers;
    private double[] distances;
    public int[] indices;
    
    public KMeanCluster(double[][] data,int groups){
        this.data=data;
        this.groups=groups;
        this.centers=new double[groups][data[0].length];
        distances=new double[groups];
        indices=new int[data.length];
        for (int i=0;i<indices.length;i++) indices[i]=9;
        int m;
       
        for (int i=0;i<this.centers.length;i++){
            centers[i]=data[i];
        }
        
    }
    
    public double euclid(double[] d1,double[] d2){
        double result=0.0;
        for (int i=0;i<d1.length;i++) result+=Math.pow(d1[i]-d2[i],2.0);
        return (Math.sqrt(result));
    }
    
    public double[] Sum(double[] d1,double[] d2){
        double[] result=new double[d1.length];
        for (int i=0;i<d1.length;i++) result[i]=d1[i]+d2[i];
        return (result);
    }
    
    public double[] Devide(double[] d1,double b){
        double[] result=new double[d1.length];
        for (int i=0;i<d1.length;i++) result[i]=d1[i]/b;
        return (result);
    }
    
    
    public void updateCenters(){
        double mindis;
        int[] ns=new int[groups];
        double distance;
        for (int i=0;i<data.length;i++){
            mindis=999999999;
            //indices[i]=-1;
            for (int j=0;j<centers.length;j++){
                distance=euclid(data[i],centers[j]);
                if (distance<mindis) {
                    //System.out.println(distance+" is smaller than "+mindis+" and obs "+i+" is belong to grp "+j );
                    mindis=distance;
                    indices[i]=j;
                }
            }
                        
        }
        
        centers=new double[groups][data[0].length];
        
        for (int i=0;i<indices.length;i++){
            centers[indices[i]]=Sum(centers[indices[i]],data[i]);
            ns[indices[i]]++;
        }
        
        for (int i=0;i<centers.length;i++){
            centers[i]=Devide(centers[i],ns[i]);
        }
    }
    
    public void printIndices(){
        for (int i=0;i<data.length;i++) System.out.print(indices[i]);
        System.out.print(" - ");
        for (int i=0;i<centers.length;i++) {
            for (int j=0;j<centers[0].length;j++){
                System.out.print(centers[i][j]+" ");
            }
            
        }
        System.out.println();
    }
    
    public void doIt(){
        for (int iter=0;iter<50;iter++){
            
            this.updateCenters();this.printIndices();
            
        }
    }
    
}