confusionmatrix.java

著名的开源仿真软件yale

/*
 *  YALE - Yet Another Learning Environment
 *  Copyright (C) 2002, 2003
 *      Simon Fischer, Ralf Klinkenberg, Ingo Mierswa, 
 *          Katharina Morik, Oliver Ritthoff
 *      Artificial Intelligence Unit
 *      Computer Science Department
 *      University of Dortmund
 *      44221 Dortmund,  Germany
 *  email: [e-mail]
 *  web:   http://yale.cs.uni-dortmund.de/
 *
 *  This program is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU General Public License as 
 *  published by the Free Software Foundation; either version 2 of the
 *  License, or (at your option) any later version. 
 *
 *  This program is distributed in the hope that it will be useful, but
 *  WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 *  General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
 *  USA.
 */
package edu.udo.cs.yale.operator.performance;

import edu.udo.cs.yale.example.ExampleSet;
import edu.udo.cs.yale.example.Example;

/** A confusion matrix.
 *  @version $Id: ConfusionMatrix.java,v 2.3 2003/04/15 19:49:47 fischer Exp $
 */
public class ConfusionMatrix extends MeasuredPerformance {
    
    private double threshold;
    private double thresholdLabel;
    private double[][] costMatrix;
    private double totalCost;
    private int exampleCount;

    /** If the (predicted) label exceeds the threshold is considered to be positive, otherwise negative.
     *  If the classification is correct, the profit is increased by pp or nn, if the label is positive
     *  and the predicted label is negative, the profit is increased by pn. In the opposite case the profit
     *  is increased by np. */
    public ConfusionMatrix(double threshold, double thresholdLabel, double pp, double pn, double np, double nn) {
	this.threshold = threshold;
	this.thresholdLabel = thresholdLabel;	
	this.costMatrix = new double[2][2];
	this.costMatrix[0][0] = pp;
	this.costMatrix[0][1] = pn;
	this.costMatrix[1][0] = np;
	this.costMatrix[1][1] = nn;
    }

    public String getName() {
	return "data mining performance";
    }

    private ConfusionMatrix(ConfusionMatrix c) {
	this.threshold = c.threshold;
	this.thresholdLabel = c.thresholdLabel;
	this.costMatrix = new double[2][2];
	this.costMatrix[0][0] = c.costMatrix[0][0];
	this.costMatrix[0][1] = c.costMatrix[0][1];
	this.costMatrix[1][0] = c.costMatrix[1][0];
	this.costMatrix[1][1] = c.costMatrix[1][1];
    }

    public void startCounting(ExampleSet set) { 
	super.startCounting(set);
	totalCost = 0.0;
    }

    public double getFitness() {
	return getValue();
    }

    public double getValue() {
	return totalCost / exampleCount;
    }

    public double getVariance() {
	return Double.NaN;
    }

    public void countExample(Example e) {
	int l = e.getLabel() >= thresholdLabel ? 0 : 1;
	int p = e.getPredictedLabel() >= threshold ? 0 : 1;
	totalCost += costMatrix[l][p];
	exampleCount++;
    }

    public int compareTo(Object o) {
	double dif = ((ConfusionMatrix)o).getValue() - getValue();
	if (dif < 0) return 1;
	if (dif > 0) return -1;
	return 0;
    }

    public Object clone() {
	return new ConfusionMatrix(this);
    }


    public double getThreshold() { return threshold; }

    public String toString() {
	return super.toString() + " (threshold="+threshold+")";
    }

    public void buildAverage(PerformanceCriterion performance) {
	super.buildAverage(performance);
	ConfusionMatrix other = (ConfusionMatrix)performance;
	this.totalCost    += other.totalCost;
	this.exampleCount += other.exampleCount;
    }
}