wekalearner.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.learner.weka;

import edu.udo.cs.yale.operator.parameter.*;
import edu.udo.cs.yale.operator.OperatorException;
import edu.udo.cs.yale.operator.UserError;
import edu.udo.cs.yale.operator.learner.AbstractLearner;
import edu.udo.cs.yale.operator.learner.Model;
import edu.udo.cs.yale.example.Attribute;
import edu.udo.cs.yale.example.Example;
import edu.udo.cs.yale.example.ExampleReader;
import edu.udo.cs.yale.example.ExampleSet;
import edu.udo.cs.yale.tools.Ontology;
import edu.udo.cs.yale.tools.LogService;
import edu.udo.cs.yale.tools.WekaTools;

import weka.classifiers.Classifier;
import weka.core.Instances;

import java.util.List;
import java.util.Iterator;

/** This operator can build all classifiers from the 
 *  <a href="http://www.cs.waikato.ac.nz/~ml/weka/">Weka</a> package.<br/>
 *  The classifier type can be selected by the parameter <var>weka_learner_name</var>. 
 *  Parameters can be passed to the Weka classifier using the Yale parameter list
 *  <var>weka_parameters</var>. The leading dash &quot;-&quot; for the keys must be omitted. 
 *  See the Weka javadoc for classifier and parameter descriptions.<br/>
 *
 *  Some Weka operators like bagging and boosting forward parameters to inner classifiers
 *  (specified by parameter W and the classname). To separate these parameters from the parameters
 *  of the wrapper, use a single parameter with key &quot;-&quot; and empty value before each 
 *  parameter value which should be passed to the learning scheme.
 *
 *  For all meta learning schemes which uses only one other weka learner (usually with option W) 
 *  the operator {@link WekaMetaLearner} can also be used.
 *
 *  @yale.xmlclass WekaLearner
 *  @version $Id: WekaLearner.java,v 1.3 2004/08/27 11:57:42 ingomierswa Exp $
 */
public class WekaLearner extends AbstractLearner {

    public static final String[] WEKA_CLASSIFIERS = WekaTools.getWekaClasses(weka.classifiers.Classifier.class);


    public Model learn(ExampleSet exampleSet) throws OperatorException {
	String operatorName = getWekaLearnerName();
	String[] parameters = getWekaParameters();

	boolean useDist = getParameterAsBoolean("use_distribution");
	if (useDist && (!exampleSet.getLabel().isNominal()))
	    throw new UserError(this, 101, "distribution learning", exampleSet.getLabel().getName());

	Classifier classifier = null;
	try {
	    classifier = Classifier.forName(operatorName, parameters);
	} catch (Exception e) {
	    throw new UserError(this, e, 904, new Object[] { operatorName, e});
	}

	LogService.logMessage(getName() + ": Converting to Weka instances.", LogService.MINIMUM);
	Instances instances = WekaTools.toWekaInstances(exampleSet, "TempInstances", exampleSet.getLabel(), true);
	try {
	    LogService.logMessage(getName() + ": Building Weka classifier.", LogService.MINIMUM);
	    classifier.buildClassifier(instances);
	} catch (Exception e) {
	    throw new UserError(this, e, 905, new Object[] {operatorName, e});
	}
	return new WekaClassifier(exampleSet.getLabel(), classifier, useDist);
    }

    public String getWekaLearnerName() {
	return getParameterAsString("weka_learner_name");
    }

    public String[] getWekaParameters() {
	List wekaParameters = getParameterList("weka_parameters");
	return WekaTools.getWekaParameters(wekaParameters);
    }

    public List getParameterTypes() {
	List types = super.getParameterTypes();
	ParameterType type = new ParameterTypeStringCategory("weka_learner_name", "The fully qualified classname of the weka classifier.", WEKA_CLASSIFIERS);
	type.setExpert(false);
	types.add(type);
	types.add(new ParameterTypeList("weka_parameters", "Parameters for the Weka classifier as described in the Weka manual.", new ParameterTypeString(null, null)));
	types.add(new ParameterTypeBoolean("use_distribution", "If set to true, the prediction of the model will not be the class, but the confidence for that class.", false));
	return types;
    }
}