PCCPredictProba.java

package put.mlc.examples.pcc;

import mulan.classifier.MultiLabelOutput;
import mulan.data.MultiLabelInstances;
import put.mlc.classifiers.pcc.PCC;
import put.mlc.classifiers.pcc.inference.Inference;
import put.mlc.classifiers.pcc.inference.montecarlo.FMeasureMaximizerInference;
import put.mlc.classifiers.pcc.inference.ExhaustiveInference;
import weka.classifiers.functions.Logistic;
import weka.core.Instances;
import weka.core.converters.ConverterUtils.DataSource;
import weka.core.Utils;


public class PCCPredict {

    public static void main(String[] args) throws Exception {
        MultiLabelInstances trainSet = new MultiLabelInstances("datasets/scene-train.arff", "datasets/scene.xml");
        MultiLabelInstances testSet = new MultiLabelInstances("datasets/scene-test.arff", "datasets/scene.xml");
        Instances testData = DataSource.read("datasets/scene-test.arff");

        Inference inf = new ExhaustiveInference();
        // Inference inf = new FMeasureMaximizerInference(100, 2018);
        PCC PCCLearner = new PCC(inf);
        Logistic log = new Logistic();
        PCCLearner.setBaseClassifier(log);

        PCCLearner.build(trainSet);

        MultiLabelOutput res;
        res = PCCLearner.makePrediction(testData.get(0));
        System.out.println(Utils.arrayToString(res.getConfidences()));

        for (int i = 0; i < testData.size(); i++) {
            res = PCCLearner.makePrediction(testData.get(i));
            System.out.println(Utils.arrayToString(res.getBipartition()));
            double[] confidences = res.getConfidences();
            System.out.print("[ ");

            for (int j = 0; j < confidences.length; j++) {
                System.out.print(String.format("%.2f, ", confidences[j]));
            }

            System.out.println("]");
            System.out.println();
        }
    }
}