Repository /RSCTC/2010/Eval_RSCTC2010.jar:rsctc2010.BalancedAccuracy

/*
 *  Copyright (C) 2009 by TunedIT
 *
 *  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 3 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, see http://www.gnu.org/licenses/.
 */
package rsctc2010;

import java.util.HashMap;
import java.util.Map;

import org.debellor.base.evaluator.score.Score;
import org.debellor.core.DataObject;


/**
 * Calculates balanced accuracy of a decision system (typically a classifier),
 * i.e. the fraction of samples for which predicted decision
 * was exactly the same as the target one,
 * calculated separately for each possible target decision
 * and averaged over target decisions, without accounting for different frequencies of decisions
 * (so rarely occuring decisions have the same impact on overall result
 * as the most frequent one).
 * 
 * @author Marcin Wojnarski
 *
 */
public class BalancedAccuracy extends Score {

	private static class Counter {
		/** Number of correct predictions for a given target decision */
		public int good = 0;
		
		/** Number of incorrect predictions for a given target decision */
		public int bad = 0;
	}

	private Map<DataObject, Counter> stats;
	
	
	public BalancedAccuracy() {
		reset();
	}

	@Override
	public void reset() {
		stats = new HashMap<DataObject, Counter>();
	}

	@Override
	public void add(DataObject target, DataObject prediction) {
		if(target == null) return;
		Counter counter = stats.get(target);
		if(counter == null) {
			counter = new Counter();
			stats.put(target, counter);
		}
		if(target.equals(prediction))
			counter.good++;
		else
			counter.bad++;
	}
	
	@Override
	public String report() {
		return "Balanced accuracy: " + result() * 100 + "%"; 
	}

	@Override
	public double result() {
		if(stats.size() == 0) return 1.0;
		
		double sum = 0.0;
		for(Counter c : stats.values())
			sum += c.good / ((double) (c.good + c.bad));
		return sum / stats.size();
	}

}