import java.util.*;

public class KMeansClustering {
   double data[][] = {{2, 3}, {1, 9}, {0, 6}, {8, 9}, {9, 6}, {6, 3}};
   double centroids[][] = {{2,3}, {1,9}};
    
   double getDistance(double[] datum, double[] centroid) {
      double d = 0.0;
      for (int i = 0; i < datum.length; i++) {
         d += Math.pow(datum[i] - centroid[i], 2); }
      return(Math.sqrt(d));
   }

   int getClosestCentroid(double[] datum) {
      double min = Double.MAX_VALUE;
      int k = -1;
      for (int i = 0; i < centroids.length; i++) {
         double d = getDistance(datum, centroids[i]);
         if (d < min) { k = i; min = d; } }
      return(k);
   }

   void printDatum(double[] datum) {
      Vector<Double> v = new Vector<Double> ();
      for (int j = 0; j < datum.length; j++) v.add(new Double(datum[j])); 
      System.out.println(v);
   } 

   void printCentroids() { 
      for (int i = 0; i < centroids.length; i++) printDatum(centroids[i]);
      System.out.println("-------------------"); 
   }

   void run(int nEpochs) {
      for (int epoch = 0; epoch < nEpochs; epoch++) {
         double newCentroids[][] = {{0,0},{0,0}};
         int n[] = {0,0}; 
         printCentroids();
         for (int i = 0; i < data.length; i++) {
            int k = getClosestCentroid(data[i]);
            newCentroids[k][0] += data[i][0];
            newCentroids[k][1] += data[i][1];
            n[k]++; } 
         newCentroids[0][0] /= n[0]; 
         newCentroids[0][1] /= n[0]; 
         newCentroids[1][0] /= n[1]; 
         newCentroids[1][1] /= n[1]; 
         centroids = newCentroids;
      }
   }

   public static void main(String args[]) {
      int n = 5; 
      if (args.length > 0) n = Integer.parseInt(args[0]);
      new KMeansClustering().run(n); 
   } 
}