---

Stats.java

gehe zur Dokumentation dieser Datei

/*
 * $Source: /shared/cvsroot/diplom/app/src/java/de/picana/math/Stats.java,v $
 * $Author: mstolpe $
 * $Date: 2003/04/22 09:51:29 $
 * $Revision: 1.3 $
 * $Release$
 *
 * Created on 21. November, 19:28
 *
 * Copyright 2002 by Marco Stolpe
 */

package de.picana.math;

import weka.core.*;


public class Stats {

    
    public static double[] getMean(Instances set) {
        
        double[] mean = new double[set.numAttributes()];

        for (int att=0; att < set.numAttributes(); att++)
            mean[att] = set.meanOrMode(att);
        
        return mean;
    }
    
    public static double getEmpVar(Instances set) {

        int num_attributes = set.numAttributes();
        double[] mean = getMean(set);
        
        double emp_var = 0.0;
        double xi_x = 0.0;
        
        for (int i=0; i < set.numInstances(); i++) {
            for (int att=0; att < num_attributes; att++) {
                xi_x = set.instance(i).value(att) - mean[att];
                emp_var += xi_x * xi_x;
            }
        }
        
        return emp_var;
    }
    
    public static double getSST(Instances set) {
        return getEmpVar(set) / set.numInstances();    
    }
    
    public static double getSSB(Instances set, Instances[] clusters) {
        
        double[] mean = getMean(set);
        double ssb = 0.0;
        double yi_y = 0.0;
        double acc_freq = 0.0;
        
        for (int cl=0; cl < clusters.length; cl++) {
            if (clusters[cl] != null) {
                acc_freq += clusters[cl].numInstances();
                double[] mean_cl = getMean(clusters[cl]);
                for (int att=0; att < mean_cl.length; att++) {
                    yi_y = mean_cl[att] - mean[att];
                    ssb += clusters[cl].numInstances() * (yi_y * yi_y);
                }
            }
        }
    
        return ssb / acc_freq;
    }
    
    public static double getSSW(Instances[] clusters) {
        
        double ssw = 0.0;
        double yi_y = 0.0;
        double acc_freq = 0.0;
        
        for (int cl=0; cl < clusters.length; cl++) {
            if (clusters[cl] != null) {
                acc_freq += clusters[cl].numInstances();
                double[] mean_cl = getMean(clusters[cl]);
                for (int i=0; i < clusters[cl].numInstances(); i++) {
                    for (int att=0; att < mean_cl.length; att++) {
                        yi_y = clusters[cl].instance(i).value(att) - mean_cl[att];
                        ssw += yi_y * yi_y;
                    }
                }
            }
        }
    
        return ssw / acc_freq;
    }
}

---