package jsat.classifiers;
import java.util.Collections;
import java.util.List;
import java.util.concurrent.ExecutorService;
import jsat.classifiers.calibration.BinaryScoreClassifier;
import jsat.parameters.Parameter;
import jsat.parameters.Parameterized;
import jsat.regression.RegressionDataSet;
import jsat.regression.Regressor;
/**
* This meta algorithm wraps a {@link Regressor} to perform binary
* classification. This is done my labeling class 0 data points as "-1" and
* class 1 points as "1". The sign of the outputs then determines the class. Not
* all regression algorithms will work well in this setting, and standard
* parameter values need to change.
* The parameter values returned are exactly those provided by the given
* regressor, or an empty list if the regressor does not implement
* {@link Parameterized}
*
* @author Edward Raff
*/
public class RegressorToClassifier implements BinaryScoreClassifier, Parameterized
{
private static final long serialVersionUID = -2607433019826385335L;
private Regressor regressor;
/**
* Creates a new Binary Classifier by using the given regressor
* @param regressor the regressor to wrap as a binary classifier
*/
public RegressorToClassifier(Regressor regressor)
{
this.regressor = regressor;
}
@Override
public double getScore(DataPoint dp)
{
return regressor.regress(dp);
}
@Override
public RegressorToClassifier clone()
{
return new RegressorToClassifier(regressor.clone());
}
@Override
public CategoricalResults classify(DataPoint data)
{
CategoricalResults cr = new CategoricalResults(2);
if(getScore(data) > 0)
cr.setProb(1, 1.0);
else
cr.setProb(0, 1.0);
return cr;
}
@Override
public void train(ClassificationDataSet dataSet, boolean parallel)
{
RegressionDataSet rds = getRegressionDataSet(dataSet);
regressor.train(rds, parallel);
}
@Override
public void train(ClassificationDataSet dataSet)
{
RegressionDataSet rds = getRegressionDataSet(dataSet);
regressor.train(rds);
}
@Override
public boolean supportsWeightedData()
{
return regressor.supportsWeightedData();
}
private RegressionDataSet getRegressionDataSet(ClassificationDataSet dataSet)
{
RegressionDataSet rds = new RegressionDataSet(dataSet.getNumNumericalVars(), dataSet.getCategories());
for(int i = 0; i < dataSet.getSampleSize(); i++)
rds.addDataPoint(dataSet.getDataPoint(i), dataSet.getDataPointCategory(i)*2-1);
return rds;
}
@Override
public List getParameters()
{
if(regressor instanceof Parameterized)
return ((Parameterized)regressor).getParameters();
else
return Collections.EMPTY_LIST;
}
@Override
public Parameter getParameter(String paramName)
{
if(regressor instanceof Parameterized)
return ((Parameterized)regressor).getParameter(paramName);
else
return null;
}
}
Ads help maintain this website.