Repository /Debellor/debellor-1.0.jar:org.debellor.base.Distortion

/*
 *  Debellor
 *
 *  Copyright (C) 2008-2009 by Marcin Wojnarski
 *
 *  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 org.debellor.base;

import org.debellor.core.Cell;
import org.debellor.core.DataObject;
import org.debellor.core.Sample;
import org.debellor.core.Sample.SampleType;
import org.debellor.core.data.DataVector;
import org.debellor.core.data.NumericFeature;
import org.debellor.core.data.DataVector.DataVectorType;
import org.debellor.core.exception.data.DataCastException;

/**
 * Introduces random distortions to numeric features of samples.
 * Each numeric feature of the data vector is disturbed independently
 * by adding random noise drawn from uniform distribution
 * on the interval <code>[-R,R]</code>,
 * where <code>R</code> is the value of the parameter "range" passed to the cell.
 * This parameter is obligatory and must have a positive real value.
 * 
 * <p>Features other than numeric are left untouched,
 * as well as the decision associated with the sample.
 * 
 * @author Marcin Wojnarski
 *
 */
public class Distortion extends Cell {

	private DataObject[] items;
	private double scale;
	private Stream input;

	public Distortion() {
		super(false);
	}

	@Override
	protected SampleType onOpen() throws Exception {
		input = openInputStream();
		SampleType type = input.sampleType;
		
		scale = 2.0 * parameters.getAsDouble("range");
		
		DataVectorType data = (DataVectorType) type.data;
		items = new DataObject[data.size()];
		
		return type;
	}

	@Override
	protected Sample onNext() throws Exception {
		Sample s = input.next();
		if(s == null) return null;
		
		DataVector inData = (DataVector) s.data;
		if(inData.size() != items.length) throw new Exception();
		
		for(int i = 0; i < items.length; i++) {
			items[i] = inData.get(i);
			try {
				NumericFeature inItem = items[i].asNumericFeature();
				double d = (random.nextDouble() - 0.5) * scale;
				items[i] = new NumericFeature(inItem.value + d);
			}
			catch(DataCastException e) {}	// if the feature is not numeric, do nothing
		}
		
		DataVector data = new DataVector(items);
		return s.setData(data);
	}

	@Override
	protected void onClose() throws Exception {
		input.close();
		items = null;
	}

}