randomsplitmethodvalidationchain.java

一个很好的LIBSVM的JAVA源码。对于要研究和改进SVM算法的学者。可以参考。来自数据挖掘工具YALE工具包。

/*
 *  YALE - Yet Another Learning Environment
 *  Copyright (C) 2001-2004
 *      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.validation;

import edu.udo.cs.yale.operator.IOObject;
import edu.udo.cs.yale.operator.IOContainer;
import edu.udo.cs.yale.operator.OperatorException;
import edu.udo.cs.yale.operator.parameter.*;
import edu.udo.cs.yale.tools.LogService;
import edu.udo.cs.yale.example.ExampleSet;
import edu.udo.cs.yale.example.SplittedExampleSet;
import edu.udo.cs.yale.operator.learner.Model;
import edu.udo.cs.yale.operator.performance.PerformanceVector;
import edu.udo.cs.yale.operator.performance.PerformanceCriterion;

import java.util.List;

/** This operator evaluates the performance of algorithms, e.g. feature selection algorithms. The first
 *  inner operator is the algorithm to be evaluated itself. It must return an example set which is in turn
 *  used to create a new model using the second inner operator and retrieve a performance vector using
 *  the third inner operator. This performance vector serves as a performance indicator for the actual algorithm.
 *
 *  This implementation the example as described for the {@link RandomSplitValidationChain}.
 *
 *  @yale.xmlclass RandomSplitMethodValidationChain  
 *  @author ingo
 *  @version $Id: RandomSplitMethodValidationChain.java,v 1.3 2004/09/14 08:39:07 ingomierswa Exp $
 */
public class RandomSplitMethodValidationChain extends MethodValidationChain {

    public IOObject[] apply() throws OperatorException {
	double splitRatio = getParameterAsDouble("split_ratio");
	IOContainer input = getInput();
	SplittedExampleSet eSet = new SplittedExampleSet((ExampleSet)input.getInput(ExampleSet.class),
							 splitRatio);

	eSet.selectSingleSubset(0);
	ExampleSet methodExampleSet = (ExampleSet)useMethod(eSet).getInput(ExampleSet.class);
	SplittedExampleSet newInputSet = (SplittedExampleSet)eSet.clone();
	newInputSet.setAttributes(methodExampleSet);

	learn(newInputSet);
	newInputSet.selectSingleSubset(1);
	IOContainer evalRes = evaluate(newInputSet);

	PerformanceVector pv = (PerformanceVector)evalRes.getInput(PerformanceVector.class);
	setResult(pv.getMainCriterion());
		
	return new IOObject[] { pv, null };
    }

    public int getNumberOfValidationSteps() {
	return 1;
    }
    
    public List getParameterTypes() {
	List types = super.getParameterTypes();
	ParameterType type = new ParameterTypeDouble("split_ratio", "Relative size of the training set.", 0, 1, 0.7);
	type.setExpert(false);
	types.add(type);
	return types;
    }
}