Java Implementation of Kmeans Algorithm on Linear Dataset ( Based on Random Input )

Kmeans.java

/** The class encapsulates an implementation of the Kmeans algorithm for Linear Dataset
 *
 * First	argument is for Number of Elements
 * Second	argument is for Start of Randomizer
 * Third	argument is for End of Randomizer
 * Fourth	argument is for Number of Clusters
 *
 * Usage with the command line :
 * >java Kmeans 7 10 20 3
 * This Means 7 tupples of data will be created
 * With each tupple having randomized values between 10 to 40
 * And 3 clusters will be formed.
 *
 * As per Kmeans Algorithm when the datasets from previous iteration belong to the same clusters in current iteration
 * The Algorithm stops and clusters are finalized
 * 
 * @author Tomrock D'souza, St. Francis Institute Of Technology, University of Mumbai, 2017
 * @copyright GNU General Public License v3
 * No reproduction in whole or part without maintaining this copyright notice
 * and imposing this condition on any subsequent users.
 */
import java.util.*;
import java.lang.Math;
class Kmeans{
	
	public static void main(String args[]){
	
	int number,numSize,numStart,numEnd,x,y=0,cSize,l,z,i,j;
	Double temp,tempa;
	
	//Processing Arguments
	numSize=Integer.parseInt(args[0]);
	numStart=Integer.parseInt(args[1]);
	numEnd=Integer.parseInt(args[2]);
	cSize=Integer.parseInt(args[3]);
	
	x=numEnd-numStart+1;
	Integer[] rand=new Integer[x];
	for(z=numStart;z<=numEnd;z++)
	{
		rand[y]=z;
		y++;
	}
	
	//Filling Array With Random Variables
	Collections.shuffle(Arrays.asList(rand));
	Integer[] array = Arrays.copyOfRange(rand, 0, numSize+1);
	
	Integer[] clustvalue=new Integer[numSize];
	Integer[] clusttemp=new Integer[numSize];
	Double[] clust=new Double[cSize];
	y=numSize/cSize;
	z=0;x=0;
	
	//Presigning Clusters And Getting Numerical Ready For First Iteration
	for(i=0;i<cSize-1;i++)
		{l=0;
			for(j=0;j<y;j++){
			clustvalue[z]=i;
			z++;
			}
		}

		for(;z<numSize;z++)
		{
			clustvalue[z]=cSize-1;
		}
		   
	System.out.println("Array\tAssigned-cluster");		
	for(i=0;i<numSize;i++){
		System.out.println(array[i]+"\t"+clustvalue[i]);
	}
	//At this stage The first assignment of clusters to the data sets is done/
	
	

	while(true){
	for(i=0;i<cSize;i++){
		l=0;
		
		for(j=0;j<numSize;j++){
		
		if(clustvalue[j]==i){
			l+=array[j];
			x++;
			}
		}
		clust[i]=(double)l/x;
		x=0;
	}
	for(i=0;i<numSize;i++)
	{ temp=9999.9;
		for (j=0;j<cSize;j++){
			tempa=Math.abs(array[i]-clust[j]);
			if(temp>tempa){
				clusttemp[i]=j;
				temp=tempa;
			}
		}
	}
	for(i=0;i<cSize;i++){
		System.out.println("Current Cluster Value="+clust[i]);
	}
	System.out.println("\n\nArray\t\tPre-clus\tCurr-Clus");
	for(i=0;i<numSize;i++){
		System.out.println(array[i]+"\t\t"+clustvalue[i]+"\t\t"+clusttemp[i]);
	}

	//Condition that matches current and previous data set
	if(!Arrays.equals(clusttemp,clustvalue)){
		clustvalue=clusttemp.clone();
	}
	else {
		for(i=0;i<cSize;i++){
			l=0;
			
			for(j=0;j<numSize;j++){
			
			if(clustvalue[j]==i){
				l+=array[j];
				x++;
				}
			}
			clust[i]=(double)l/x;
			x=0;
		}
		System.out.println("\nFinal Cluster Values:");
		for(i=0;i<cSize;i++){
			System.out.println("Cluster Value "+i+"="+clust[i]);
		}
		break;}

	}

	}
}