net.sf.jclal.evaluation.method.HoldOut.java Source code

Introduction

Here is the source code for net.sf.jclal.evaluation.method.HoldOut.java

Source

/*
 * 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 net.sf.jclal.evaluation.method;

import java.util.logging.Level;
import java.util.logging.Logger;
import net.sf.jclal.activelearning.algorithm.AbstractALAlgorithm;
import net.sf.jclal.core.IAlgorithm;
import net.sf.jclal.dataset.AbstractDataset;
import net.sf.jclal.dataset.MulanDataset;
import net.sf.jclal.dataset.WekaDataset;
import net.sf.jclal.sampling.AbstractSampling;
import net.sf.jclal.util.dataset.DatasetUtils;
import org.apache.commons.configuration.Configuration;
import org.apache.commons.configuration.ConfigurationRuntimeException;

/**
 * Hold Out evaluation method.
 *
 * @author Oscar Gabriel Reyes Pupo
 * @author Eduardo Perez Perdomo
 */
public class HoldOut extends AbstractEvaluationMethod {

    private static final long serialVersionUID = -6193747210498415157L;
    /**
     * Percentage of instances to run the training, the rest is test set
     */
    private double percentageToSplit = 66;

    /**
     *
     * @param algorithm The algorithm.
     * @param dataset The dataset.
     * @param percentageToTrain The percentage of instances to train.
     */
    public HoldOut(AbstractALAlgorithm algorithm, AbstractDataset dataset, double percentageToTrain) {

        super(algorithm, dataset);

        this.percentageToSplit = percentageToTrain;
    }

    /**
     * Empty(default) constructor.
     */
    public HoldOut() {
    }

    /**
     *
     * @return The percentage selected to train.
     */
    public double getPercentageToSplit() {
        return percentageToSplit;
    }

    /**
     *
     * @param percent The percentage selected to train.
     */
    public void setPercentageToSplit(double percent) {
        this.percentageToSplit = percent;
    }

    /**
     * Executes the process of evaluation of the experiment
     */
    @Override
    public void evaluate() {

        try {
            //Load the initial data
            loadData();

            //One general dataset was specified, then it must be divided into train and test taking 
            //into account the percentage defined
            if (getDataset() != null) {

                DatasetUtils.randomize(((AbstractSampling) getSamplingStrategy()).getRandgen(), getDataset());

                int numInstances = (int) (getDataset().getNumInstances() * percentageToSplit / 100);

                if (!isMultiLabel()) {

                    setTrainDataset(new WekaDataset(getDataset(), 0, numInstances));
                    setTestDataset(new WekaDataset(getDataset(), numInstances,
                            getDataset().getNumInstances() - numInstances));

                } else {

                    setTrainDataset(new MulanDataset(getDataset(), 0, numInstances));
                    setTestDataset(new MulanDataset(getDataset(), numInstances,
                            getDataset().getNumInstances() - numInstances));
                }

                //Resample the instances to construct the labeled and unlabeled set on the training set
                getSamplingStrategy().sampling(getTrainDataset());

                setLabeledDataset(((AbstractSampling) getSamplingStrategy()).getLabeledData());
                setUnlabeledDataset(((AbstractSampling) getSamplingStrategy()).getUnlabeledData());

            } //confirm if a training and test file were specified
            else {

                //A training dataset was specified
                if (getTrainDataset() != null) {

                    //Resample the instances to construct the labeled and unlabeled set
                    getSamplingStrategy().sampling(getTrainDataset());

                    setLabeledDataset(((AbstractSampling) getSamplingStrategy()).getLabeledData());
                    setUnlabeledDataset(((AbstractSampling) getSamplingStrategy()).getUnlabeledData());
                }

                //Else, An unlabeled and labeled file were specified, then a resample method is not necessary
            }

            IAlgorithm algorithmCopy = getAlgorithm().makeCopy();

            algorithmCopy.setLabeledDataSet(getLabeledDataset());
            algorithmCopy.setUnlabeledDataSet(getUnlabeledDataset());
            algorithmCopy.setTestDataSet(getTestDataset());

            //Executes the algorithm
            algorithmCopy.execute();

            setFinalEvaluations(algorithmCopy.getScenario().getQueryStrategy().getEvaluations());

        } catch (Exception e) {
            Logger.getLogger(HoldOut.class.getName()).log(Level.SEVERE, null, e);
        }

    }

    /**
     * @param configuration The configuration of Hold Out.
     *
     *The XML labels supported are:
     *
     * <ul>
     * <li><b>percentage-split= double</b></li>
     * </ul>
     */
    @Override
    public void configure(Configuration configuration) {

        super.configure(configuration);

        // the percent of instances used to train
        double percentTrain = configuration.getDouble("percentage-split", percentageToSplit);

        String perc = "\n<percentage-split>" + percentTrain + "</percentage-split>";

        if (percentTrain <= 0) {
            throw new ConfigurationRuntimeException(perc + ". percentage-split > 0");
        }
        if (percentTrain >= 100) {
            throw new ConfigurationRuntimeException(perc + ". percentage-split < 100");
        }

        setPercentageToSplit(percentTrain);
    }

}