weka.clusterers.HierarchicalClusterer Java Examples

public static void main(String[] args) {
    Instances dataset = load(DATA);
    HierarchicalClusterer hc = new HierarchicalClusterer();
    hc.setLinkType(new SelectedTag(4, TAGS_LINK_TYPE));  // CENTROID
    hc.setNumClusters(3);
    try {
        hc.buildClusterer(dataset);
        for (Instance instance : dataset) {
            System.out.printf("(%.0f,%.0f): %s%n", 
                    instance.value(0), instance.value(1), 
                    hc.clusterInstance(instance));
        }
    } catch (Exception e) {
        System.err.println(e);
    }
}
 

public static void main(String[] args) {
    Instances dataset = load(DATA);
    HierarchicalClusterer hc = new HierarchicalClusterer();
    hc.setLinkType(new SelectedTag(4, TAGS_LINK_TYPE));  // CENTROID
    hc.setNumClusters(1);
    try {
        hc.buildClusterer(dataset);
        for (Instance instance : dataset) {
            System.out.printf("(%.0f,%.0f): %s%n", 
                    instance.value(0), instance.value(1), 
                    hc.clusterInstance(instance));
        }
        displayDendrogram(hc.graph());
    } catch (Exception e) {
        System.err.println(e);
    }
}
 

/**
 * Run WEKA Hierarchical clustering on the parameter ARFF file searching for
 * numClusters clusters
 * 
 * @param filename
 * @param numClusters
 * @param linkage
 * @return
 * @throws Exception
 */
public static LinkedHashMap<Integer, LinkedList<String>> runHierarchical(String filename, String numClusters,
		String linkage) throws Exception {

	String[] options = new String[6];
	options[0] = "-t";
	options[1] = filename;
	options[2] = "-N";
	options[3] = numClusters;
	options[4] = "-L";
	options[5] = linkage;

	HierarchicalClusterer c = new HierarchicalClusterer();

	c.setNumClusters(Integer.parseInt(numClusters));
	c.setDebug(false);
	c.setPrintNewick(true);

	BufferedReader reader = new BufferedReader(new FileReader(filename));
	ArffReader arff = new ArffReader(reader);
	Instances data = arff.getData();
	data.setClassIndex(0);

	c.buildClusterer(data);

	LinkedHashMap<Integer, LinkedList<String>> output = new LinkedHashMap<Integer, LinkedList<String>>();

	// initialize clusters map
	for (int i = 0; i < Integer.parseInt(numClusters); i++) {
		output.put(new Integer(i), new LinkedList<String>());
	}

	for (Instance instance : data) {
		// System.out.println(instance.stringValue(0) + "\t" +
		// c.clusterInstance(instance));
		output.get(c.clusterInstance(instance)).add(instance.stringValue(0));
	}

	return output;
}
 

/**
 * Cluster API call sequences as described in MAPO
 *
 * @return Multimap of cluster IDs to API call sequences
 */
public static Multimap<Integer, String> clusterAPICallSeqs(final String arffFile, final double threshold)
		throws Exception {

	// Clusterer settings
	final HierarchicalClusterer clusterer = new HierarchicalClusterer();
	clusterer.setOptions(new String[] { "-L", "COMPLETE" }); // link type
	clusterer.setDebug(true);
	clusterer.setNumClusters(1);
	clusterer.setDistanceFunction(MAPOSimilarity);
	clusterer.setDistanceIsBranchLength(false);

	// Read in API call seqs
	final DataSource source = new DataSource(arffFile);
	final Instances data = source.getDataSet();

	// Cluster API call seqs
	clusterer.buildClusterer(data);

	// Assign seqs to clusters based on dendrogram
	final String newick = clusterer.graph().replace("Newick:", "") + ":0";
	if (newick.equals("(no,clusters):0")) // Handle no clusters
		return HashMultimap.create();
	final Multimap<Integer, String> clusters = NewickTreeParser.getClusters(newick, threshold);
	System.out.println("No. clusters: " + clusters.keySet().size());
	final Multimap<Integer, String> assignments = HashMultimap.create();
	for (int i = 0; i < data.numInstances(); i++) {
		for (final int id : clusters.keySet()) {
			if (clusters.get(id).contains(data.instance(i).stringValue(0)))
				assignments.put(id, data.instance(i).stringValue(1));
		}
	}

	// showDendrogram(clusterer);

	return assignments;
}
 

public static void showDendrogram(final HierarchicalClusterer clusterer) throws Exception {
	final JFrame mainFrame = new JFrame("Dendrogram");
	mainFrame.setSize(1024, 768);
	mainFrame.setDefaultCloseOperation(JFrame.DISPOSE_ON_CLOSE);
	final Container content = mainFrame.getContentPane();
	content.setLayout(new GridLayout(1, 1));

	final HierarchyVisualizer visualizer = new HierarchyVisualizer(clusterer.graph());
	content.add(visualizer);

	mainFrame.setVisible(true);
}
 

/**
 * Cluster API call sequences as described in UPMiner
 *
 * @return Multimap of cluster IDs to API call sequences
 */
public static Multimap<Integer, String> clusterAPICallSeqs(final String arffFile, final double threshold)
		throws Exception {

	// Clusterer settings
	final HierarchicalClusterer clusterer = new HierarchicalClusterer();
	clusterer.setOptions(new String[] { "-L", "COMPLETE" }); // link type
	clusterer.setDebug(true);
	clusterer.setNumClusters(1);
	clusterer.setDistanceFunction(SeqSimilarity);
	clusterer.setDistanceIsBranchLength(false);

	// Read in API call seqs
	final DataSource source = new DataSource(arffFile);
	final Instances data = source.getDataSet();

	// Cluster API call seqs
	clusterer.buildClusterer(data);

	// Assign seqs to clusters based on dendrogram
	final String newick = clusterer.graph().replace("Newick:", "") + ":0";
	if (newick.equals("(no,clusters):0")) // Handle no clusters
		return HashMultimap.create();
	final Multimap<Integer, String> clusters = NewickTreeParser.getClusters(newick, threshold);
	System.out.println("No. clusters: " + clusters.keySet().size());
	final Multimap<Integer, String> assignments = HashMultimap.create();
	for (int i = 0; i < data.numInstances(); i++) {
		for (final int id : clusters.keySet()) {
			if (clusters.get(id).contains(data.instance(i).stringValue(0)))
				assignments.put(id, data.instance(i).stringValue(1));
		}
	}

	// showDendrogram(clusterer);

	return assignments;
}
 

public static void showDendrogram(final HierarchicalClusterer clusterer) throws Exception {
	final String timeStamp = new SimpleDateFormat("yyyy/MM/dd HH:mm:ss").format(new Date());
	final JFrame mainFrame = new JFrame("Dendrogram " + timeStamp);
	mainFrame.setSize(1024, 768);
	mainFrame.setDefaultCloseOperation(JFrame.DISPOSE_ON_CLOSE);
	final Container content = mainFrame.getContentPane();
	content.setLayout(new GridLayout(1, 1));

	final HierarchyVisualizer visualizer = new HierarchyVisualizer(clusterer.graph());
	content.add(visualizer);

	mainFrame.setVisible(true);
}
 

@Override
public ClusteringResult performClustering(Instances dataset, ParameterSet parameters) {
  HierarchicalClusterer clusterer = new HierarchicalClusterer();
  String[] options = new String[5];
  LinkType link = parameters.getParameter(HierarClustererParameters.linkType).getValue();
  DistanceType distanceType =
      parameters.getParameter(HierarClustererParameters.distanceType).getValue();
  options[0] = "-L";
  options[1] = link.name();
  options[2] = "-A";
  switch (distanceType) {
    case EUCLIDIAN:
      options[3] = "weka.core.EuclideanDistance";
      break;
    case CHEBYSHEV:
      options[3] = "weka.core.ChebyshevDistance";
      break;
    case MANHATTAN:
      options[3] = "weka.core.ManhattanDistance";
      break;
    case MINKOWSKI:
      options[3] = "weka.core.MinkowskiDistance";
      break;
  }

  options[4] = "-P";
  try {
    clusterer.setOptions(options);
    clusterer.setPrintNewick(true);
    clusterer.buildClusterer(dataset);
    // clusterer.graph() gives only the first cluster and in the case
    // there
    // are more than one cluster the variables in the second cluster are
    // missing.
    // I'm using clusterer.toString() which contains all the clusters in
    // Newick format.
    ClusteringResult result =
        new ClusteringResult(null, clusterer.toString(), clusterer.getNumClusters(), null);
    return result;
  } catch (Exception ex) {
    logger.log(Level.SEVERE, null, ex);
    return null;
  }
}
 

@Override
public ClusteringResult performClustering(Instances dataset, ParameterSet parameters) {
  HierarchicalClusterer clusterer = new HierarchicalClusterer();
  String[] options = new String[5];
  LinkType link = parameters.getParameter(HierarClustererParameters.linkType).getValue();
  DistanceType distanceType =
      parameters.getParameter(HierarClustererParameters.distanceType).getValue();
  options[0] = "-L";
  options[1] = link.name();
  options[2] = "-A";
  switch (distanceType) {
    case EUCLIDIAN:
      options[3] = "weka.core.EuclideanDistance";
      break;
    case CHEBYSHEV:
      options[3] = "weka.core.ChebyshevDistance";
      break;
    case MANHATTAN:
      options[3] = "weka.core.ManhattanDistance";
      break;
    case MINKOWSKI:
      options[3] = "weka.core.MinkowskiDistance";
      break;
  }

  options[4] = "-P";
  try {
    clusterer.setOptions(options);
    clusterer.setPrintNewick(true);
    clusterer.buildClusterer(dataset);
    // clusterer.graph() gives only the first cluster and in the case
    // there
    // are more than one cluster the variables in the second cluster are
    // missing.
    // I'm using clusterer.toString() which contains all the clusters in
    // Newick format.
    ClusteringResult result =
        new ClusteringResult(null, clusterer.toString(), clusterer.getNumClusters(), null);
    return result;
  } catch (Exception ex) {
    logger.log(Level.SEVERE, null, ex);
    return null;
  }
}