1. public double[][] sampleGibbsClassificationsIG(double[] observations, double[][] parameters) {
  2.  
  3.         Random ran = new Random();
  4.         double[][] observationClass = new double[observations.length][2];
  5.         Double[][] probabilities = new Double[parameters.length][2];
  6.         double[] IGparams = new double[2];
  7.         Double constant = 0.0;
  8.         double random = 0.0;
  9.  
  10.  
  11.         for (int i = 0; i < observations.length; i++) {
  12.  
  13.             random = ran.nextDouble();
  14.             observationClass[i][0] = observations[i];
  15.  
  16.             for (int k = 0; k < parameters.length; k++) {
  17.  
  18.                 IGparams[0] = parameters[k][0];
  19.                 IGparams[1] = parameters[k][1];
  20.                 probabilities[k][0] = (Double) (Distribution.computeIGDensityGeneral(observations[i], IGparams) * parameters[k][2]);
  21.                 constant += probabilities[k][0];
  22.                 //System.out.println("Probabilities: " + probabilities[k][0]);
  23.             }
  24.             //Comparator sorts the fields in descending order
  25.             Arrays.sort(probabilities, new DoubleArrayComparator(0));
  26.  
  27.             //Look now if the argumentum minimum over the sequence of the sum of the probabilities ist smaller than the random draw.
  28.             for (int k = 0; k < parameters.length; k++) {
  29.  
  30.                 if (k == 0) {
  31.                     probabilities[k][1] = probabilities[k][0] / constant;
  32.                 } else {
  33.                     probabilities[k][1] = probabilities[k - 1][1] + probabilities[k][0]/constant;
  34.                 }
  35.                 //System.out.println("Probability: " + probabilities[k][1] + " random number: " + random);
  36.                 if(probabilities[k][1].compareTo((Double) random) >= 0) {
  37.                     observationClass[i][1] = k;
  38.                     break;
  39.                 }
  40.                 else continue;
  41.  
  42.             }
  43.  
  44.             constant = 0.0;
  45.         }
  46.  
  47.         return observationClass;
  48.  
  49.  
  50.  
  51.     }