/*
* 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., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
/*
* ICTAI2010.java
* Copyright (C) 2009-2010 Aristotle University of Thessaloniki, Thessaloniki, Greece
*/
package mulan.experiments;
/**
* @author Grigorios Tsoumakas
*/
import java.util.ArrayList;
import java.util.List;
import mulan.classifier.MultiLabelLearner;
import mulan.classifier.lazy.MLkNN;
import mulan.classifier.meta.thresholding.MetaLabeler;
import mulan.classifier.meta.thresholding.OneThreshold;
import mulan.classifier.meta.thresholding.RCut;
import mulan.classifier.meta.thresholding.SCut;
import mulan.classifier.meta.thresholding.ThresholdPrediction;
import mulan.classifier.neural.BPMLL;
import mulan.classifier.transformation.BinaryRelevance;
import mulan.classifier.transformation.CalibratedLabelRanking;
import mulan.data.MultiLabelInstances;
import mulan.evaluation.Evaluator;
import mulan.evaluation.MultipleEvaluation;
import mulan.evaluation.measure.HammingLoss;
import mulan.evaluation.measure.Measure;
import weka.classifiers.meta.Bagging;
import weka.classifiers.trees.J48;
import weka.classifiers.trees.M5P;
import weka.core.TechnicalInformation;
import weka.core.Utils;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;
/**
* Class replicating an experiment from a published paper
*
* @author Grigorios Tsoumakas
* @version 2010.12.10
*/
public class ICTAI2010 extends Experiment {
/**
* Main class
*
* @param args command line arguments
*/
public static void main(String[] args) {
try {
String path = Utils.getOption("path", args);
String filestem = Utils.getOption("filestem", args);
System.out.println("Loading the data set");
MultiLabelInstances dataset = new MultiLabelInstances(path + filestem + ".arff", path + filestem + ".xml");
Evaluator eval = new Evaluator();
MultipleEvaluation results;
List<Measure> measures = new ArrayList<Measure>(1);
measures.add(new HammingLoss());
int numFolds = 10;
MultiLabelLearner[] learner = new MultiLabelLearner[4];
String[] learnerName = new String[learner.length];
learner[0] = new MLkNN(10, 1.0);
learnerName[0] = "MLkNN";
learner[1] = new CalibratedLabelRanking(new J48());
learnerName[1] = "CLR";
Bagging bagging = new Bagging();
bagging.setClassifier(new J48());
learner[2] = new BinaryRelevance(bagging);
learnerName[2] = "BR";
learner[3] = new BPMLL();
learnerName[3] = "BPMLL";
// loop over learners
for (int i = 0; i < learner.length; i++) {
// Default
results = eval.crossValidate(learner[i].makeCopy(), dataset, measures, numFolds);
System.out.println(learnerName[i] + ";default;-;" + results.toCSV());
// One Threshold
OneThreshold ot;
ot = new OneThreshold(learner[i].makeCopy(), new HammingLoss());
results = eval.crossValidate(ot, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";one threshold;train;" + results.toCSV());
ot = new OneThreshold(learner[i].makeCopy(), new HammingLoss(), 5);
results = eval.crossValidate(ot, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";one threshold;5-cv;" + results.toCSV());
// RCut
RCut rcut;
rcut = new RCut(learner[i].makeCopy());
results = eval.crossValidate(rcut, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";rcut;cardinality;" + results.toCSV());
rcut = new RCut(learner[i].makeCopy(), new HammingLoss());
results = eval.crossValidate(rcut, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";rcut;train;" + results.toCSV());
rcut = new RCut(learner[i].makeCopy(), new HammingLoss(), 5);
results = eval.crossValidate(rcut, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";rcut;5-cv;" + results.toCSV());
// SCut
SCut scut;
scut = new SCut(learner[i].makeCopy(), new HammingLoss());
results = eval.crossValidate(scut, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";scut;train;" + results.toCSV());
scut = new SCut(learner[i].makeCopy(), new HammingLoss(), 5);
results = eval.crossValidate(scut, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";scut;5-cv;" + results.toCSV());
// MetaLabeler
MetaLabeler ml;
ml = new MetaLabeler(learner[i].makeCopy(), new M5P(), "Content-Based", "Numeric-Class");
ml.setFolds(1);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;m5p;train;content;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new M5P(), "Score-Based", "Numeric-Class");
ml.setFolds(1);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;m5p;train;scores;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new M5P(), "Rank-Based", "Numeric-Class");
ml.setFolds(1);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;m5p;train;ranks;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new J48(), "Content-Based", "Nominal-Class");
ml.setFolds(1);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;j48;train;content;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new J48(), "Score-Based", "Nominal-Class");
ml.setFolds(1);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;j48;train;scores;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new J48(), "Rank-Based", "Nominal-Class");
ml.setFolds(1);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;j48;cv;ranks;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new M5P(), "Content-Based", "Numeric-Class");
ml.setFolds(5);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;m5p;cv;content;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new M5P(), "Score-Based", "Numeric-Class");
ml.setFolds(5);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;m5p;cv;scores;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new M5P(), "Rank-Based", "Numeric-Class");
ml.setFolds(5);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;m5p;cv;ranks;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new J48(), "Content-Based", "Nominal-Class");
ml.setFolds(5);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;j48;cv;content;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new J48(), "Score-Based", "Nominal-Class");
ml.setFolds(5);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;j48;cv;scores;" + results.toCSV());
ml = new MetaLabeler(learner[i].makeCopy(), new J48(), "Rank-Based", "Nominal-Class");
ml.setFolds(5);
results = eval.crossValidate(ml, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";metalabeler;j48;cv;ranks;" + results.toCSV());
// ThresholdPrediction
ThresholdPrediction tp;
tp = new ThresholdPrediction(learner[i].makeCopy(), new M5P(), "Content-Based", 1);
results = eval.crossValidate(tp, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";tp;m5p;train;content;" + results.toCSV());
tp = new ThresholdPrediction(learner[i].makeCopy(), new M5P(), "Score-Based", 1);
results = eval.crossValidate(tp, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";tp;m5p;train;scores;" + results.toCSV());
tp = new ThresholdPrediction(learner[i].makeCopy(), new M5P(), "Rank-Based", 1);
results = eval.crossValidate(tp, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";tp;m5p;train;ranks;" + results.toCSV());
tp = new ThresholdPrediction(learner[i].makeCopy(), new M5P(), "Content-Based", 5);
results = eval.crossValidate(tp, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";tp;m5p;5-cv;content;" + results.toCSV());
tp = new ThresholdPrediction(learner[i].makeCopy(), new M5P(), "Score-Based", 5);
results = eval.crossValidate(tp, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";tp;m5p;5-cv;scores;" + results.toCSV());
tp = new ThresholdPrediction(learner[i].makeCopy(), new M5P(), "Rank-Based", 5);
results = eval.crossValidate(tp, dataset, measures, numFolds);
System.out.println(learnerName[i] + ";tp;m5p;5-cv;ranks;" + results.toCSV());
}
} catch (Exception e) {
e.printStackTrace();
}
}
/**
* Returns an instance of a TechnicalInformation object, containing detailed
* information about the technical background of this class, e.g., paper
* reference or book this class is based on.
*
* @return the technical information about this class
*/
@Override
public TechnicalInformation getTechnicalInformation() {
TechnicalInformation result;
result = new TechnicalInformation(Type.INPROCEEDINGS);
result.setValue(Field.AUTHOR, "Marios Ioannou and George Sakkas and Grigorios Tsoumakas and Ioannis Vlahavas");
result.setValue(Field.TITLE, "Obtaining Bipartitions from Score Vectors for Multi-Label Classification");
result.setValue(Field.BOOKTITLE, "Proceedings of the 22th IEEE International Conference on Tools with Artificial Intelligence");
result.setValue(Field.YEAR, "2010");
result.setValue(Field.PUBLISHER, "IEEE");
return result;
}
}