private static IList<IList<IAgent>> clusteringUsingWeka(
final IScope scope,
final Clusterer clusterer,
final IList<String> attributes,
final IAddressableContainer<Integer, IAgent, Integer, IAgent> agents)
throws GamaRuntimeException {
Instances dataset = convertToInstances(scope, attributes, agents);
try {
clusterer.buildClusterer(dataset);
IList<IList<IAgent>> groupes = GamaListFactory.create(Types.LIST.of(Types.AGENT));
for (int i = 0; i < clusterer.numberOfClusters(); i++) {
groupes.add(GamaListFactory.<IAgent>create(Types.AGENT));
}
for (int i = 0; i < dataset.numInstances(); i++) {
Instance inst = dataset.instance(i);
int clusterIndex = -1;
clusterIndex = clusterer.clusterInstance(inst);
IList<IAgent> groupe = groupes.get(clusterIndex);
groupe.add(agents.get(scope, i));
}
return groupes;
} catch (Exception e) {
return null;
}
}
/**
* Set the clusterer for this wrapper
*
* @param c a <code>weka.clusterers.Clusterer</code> value
*/
public void setClusterer(weka.clusterers.Clusterer c) {
boolean loadImages = true;
if (c.getClass().getName().compareTo(m_Clusterer.getClass().getName()) == 0) {
loadImages = false;
} else {
// clusterer has changed so any batch training status is now
// invalid
m_trainingSet = null;
}
m_Clusterer = c;
String clustererName = c.getClass().toString();
clustererName =
clustererName.substring(clustererName.lastIndexOf('.') + 1, clustererName.length());
if (loadImages) {
if (!m_visual.loadIcons(
BeanVisual.ICON_PATH + clustererName + ".gif",
BeanVisual.ICON_PATH + clustererName + "_animated.gif")) {
useDefaultVisual();
}
}
m_visual.setText(clustererName);
// get global info
m_globalInfo = KnowledgeFlowApp.getGlobalInfo(m_Clusterer);
}
/** Builds the clusters */
private void buildClusterer() throws Exception {
if (m_trainingSet.classIndex() < 0) m_Clusterer.buildClusterer(m_trainingSet);
else { // class based evaluation if class attribute is set
Remove removeClass = new Remove();
removeClass.setAttributeIndices("" + (m_trainingSet.classIndex() + 1));
removeClass.setInvertSelection(false);
removeClass.setInputFormat(m_trainingSet);
Instances clusterTrain = Filter.useFilter(m_trainingSet, removeClass);
m_Clusterer.buildClusterer(clusterTrain);
}
}
/**
* Zwraca tylko te instancje, kt??re nale???? do klastra o podanym numerze.
*
* @param numClust Numer klastra
* @return Instancje nale????ce do podanego klastra.
*/
public Instances getClusterInstances(int numClust) {
Instances inst = new Instances(data);
inst.clear();
try {
for (int i = 0; i < data.numInstances(); ++i) {
if (clusterer.clusterInstance(data.instance(i)) == numClust) {
inst.add(data.instance(i));
}
}
} catch (Exception e) {
e.printStackTrace();
}
return inst;
}
@Override
public List<String> getPredictionLabels() throws DistributedWekaException {
if (m_predictionLabels == null) {
m_predictionLabels = new ArrayList<String>();
try {
for (int i = 0; i < m_model.numberOfClusters(); i++) {
m_predictionLabels.add("Cluster_" + i);
}
} catch (Exception ex) {
throw new DistributedWekaException(ex);
}
}
return m_predictionLabels;
}
/**
* Get the number of clusters from the base clusterer
*
* @return the number of clusters
* @throws Exception if a problem occurs
*/
public int numberOfClusters() throws Exception {
return m_model.numberOfClusters();
}
@Override
public double[] distributionForInstance(Instance inst) throws Exception {
return m_model.distributionForInstance(inst);
}
