/**
* Generates a clusterer.
*
* @param data the training instances
* @throws Exception if something goes wrong
*/
@Override
public void buildClusterer(Instances data) throws Exception {
// can clusterer handle the data ?
getCapabilities().testWithFail(data);
m_replaceMissing = new ReplaceMissingValues();
Instances instances = new Instances(data);
instances.setClassIndex(-1);
m_replaceMissing.setInputFormat(instances);
data = weka.filters.Filter.useFilter(instances, m_replaceMissing);
instances = null;
// initialize all fields that are not being set via options
m_data = data;
m_numInstances = m_data.numInstances();
m_numAttributes = m_data.numAttributes();
random = new Random(getSeed());
// initialize the statistics of the input training data
input = sIB_ProcessInput();
// object to hold the best partition
bestT = new Partition();
// the real clustering
double bestL = Double.NEGATIVE_INFINITY;
for (int k = 0; k < m_numRestarts; k++) {
if (m_verbose) {
System.out.format("restart number %s...\n", k);
}
// initialize the partition and optimize it
Partition tmpT = sIB_InitT(input);
tmpT = sIB_OptimizeT(tmpT, input);
// if a better partition is found, save it
if (tmpT.L > bestL) {
tmpT.copy(bestT);
bestL = bestT.L;
}
if (m_verbose) {
System.out.println("\nPartition status : ");
System.out.println("------------------");
System.out.println(tmpT.toString() + "\n");
}
}
if (m_verbose) {
System.out.println("\nBest Partition");
System.out.println("===============");
System.out.println(bestT.toString());
}
// save memory
m_data = new Instances(m_data, 0);
}
/**
* Adds Partition to and validates Partition is unique. A duplicate Partition results in an
* exception.
*/
private void addPartition(Partition partition) {
if (partitionMap.put(ByteBuffer.wrap(partition.getBytes()), partition) != null) {
throw new IllegalStateException("Duplicate Partition detected: " + partition.toString());
}
if (partition.getId() >= maxPartitionId) {
maxPartitionId = partition.getId() + 1;
}
}