Class MajorityVote

JSAT/src/jsat/classifiers/MajorityVote.java:14–104

The Majority Vote classifier is a simple ensemble classifier. Given a list of base classifiers, it will sum the most likely votes from each base classifier and return a result based on the majority votes. It does not take into account the confidence of the votes. @author Edward Raff

 * @author Edward Raff
 */
public class MajorityVote implements Classifier
{

	private static final long serialVersionUID = 7945429768861275845L;
	private Classifier[] voters;

    /**
     * Creates a new Majority Vote classifier using the given voters. If already trained, the 
     * Majority Vote classifier can be used immediately. The MajorityVote does not make 
     * copies of the given classifiers. <br>
     * <tt>null</tt> values in the array will have no vote. 
     * 
     * @param voters the array of voters to use
     */
    public MajorityVote(Classifier... voters)
    {
        this.voters = voters;
    }
    
    /**
     * Creates a new Majority Vote classifier using the given voters. If already trained, the 
     * Majority Vote classifier can be used immediately. The MajorityVote does not make 
     * copies of the given classifiers. <br>
     * <tt>null</tt> values in the array will have no vote. 
     * 
     * @param voters the list of voters to use
     */
    public MajorityVote(List<Classifier> voters)
    {
        this.voters = voters.toArray(new Classifier[0]);
    }
    
    @Override
    public CategoricalResults classify(DataPoint data)
    {
        CategoricalResults toReturn = null;

        for (Classifier classifier : voters)
            if (classifier != null)
                if (toReturn == null)
                {
                    toReturn = classifier.classify(data);
                    //Instead of allocating a new catResult, reuse the given one. Set the non likely to zero, and most to 1. 
                    for (int i = 0; i < toReturn.size(); i++)
                        if (i != toReturn.mostLikely())
                            toReturn.setProb(i, 0);
                        else
                            toReturn.setProb(i, 1.0);
                }
                else
                {
                    CategoricalResults vote = classifier.classify(data);
                    for (int i = 0; i < toReturn.size(); i++)
                        toReturn.incProb(vote.mostLikely(), 1.0);
                }

        toReturn.normalize();
        return toReturn;