/**
* Normalize the weights for the next iteration.
*
* @param training the training instances
* @throws Exception if something goes wrong
*/
protected void normalizeWeights(Instances training, double oldSumOfWeights) throws Exception {
// Renormalize weights
double newSumOfWeights = training.sumOfWeights();
for (Instance instance : training) {
instance.setWeight(instance.weight() * oldSumOfWeights / newSumOfWeights);
}
}
/**
* Sets the weights for the next iteration.
*
* @param training the data to train with
* @param reweight the reweighting factor
* @throws Exception in case of an error
*/
protected void setWeights(Instances training, double reweight) throws Exception {
int subCmtySize = m_Classifiers.length / m_NumSubCmtys;
if ((m_NumIterationsPerformed + 1) % subCmtySize == 0) {
if (getDebug()) System.err.println(m_NumIterationsPerformed + " " + subCmtySize);
double oldSumOfWeights = training.sumOfWeights();
// Randomly set the weights of the training instances to the poisson distributon
for (int i = 0; i < training.numInstances(); i++) {
training.instance(i).setWeight(-Math.log((m_Random.nextDouble() * 9999) / 10000));
}
// Renormailise weights
double sumProbs = training.sumOfWeights();
for (int i = 0; i < training.numInstances(); i++) {
training.instance(i).setWeight(training.instance(i).weight() * oldSumOfWeights / sumProbs);
}
} else {
super.setWeights(training, reweight);
}
}
private void newDistribution(Instances data) throws Exception {
Instances[] localInstances;
m_localModel.resetDistribution(data);
m_train = data;
if (!m_isLeaf) {
localInstances = (Instances[]) m_localModel.split(data);
for (int i = 0; i < m_sons.length; i++) {
m_sons[i].newDistribution(localInstances[i]);
}
} else {
// Check whether there are some instances at the leaf now!
if (!Utils.eq(data.sumOfWeights(), 0)) {
m_isEmpty = false;
}
}
}
/**
* Boosting method.
*
* @param data the training data to be used for generating the boosted classifier.
* @throws Exception if the classifier could not be built successfully
*/
public void buildClassifier(Instances data) throws Exception {
super.buildClassifier(data);
// can classifier handle the data?
getCapabilities().testWithFail(data);
// remove instances with missing class
data = new Instances(data);
data.deleteWithMissingClass();
m_SumOfWeights = data.sumOfWeights();
if ((!m_UseResampling) && (m_Classifier instanceof WeightedInstancesHandler)) {
buildClassifierWithWeights(data);
} else {
buildClassifierUsingResampling(data);
}
}
public void buildTree(Instances data) throws Exception {
Instances[] localInstances;
// m_test = null;
m_isLeaf = false;
m_isEmpty = false;
m_sons = null;
m_localModel = m_toSelectModel.selectModel(data);
if (m_localModel.numSubsets() > 1) {
localInstances = m_localModel.split(data);
data = null;
m_sons = new myJ48ClassifierTree[m_localModel.numSubsets()];
for (int i = 0; i < m_sons.length; i++) {
m_sons[i] = getNewTree(localInstances[i]);
localInstances[i] = null;
}
} else {
m_isLeaf = true;
if (Utils.eq(data.sumOfWeights(), 0)) {
m_isEmpty = true;
}
data = null;
}
}
/**
* Compute the combined DL of the ruleset in this class, i.e. theory DL and data DL. Note this
* procedure computes the combined DL according to the current status of the ruleset in this class
*
* @param expFPRate expected FP/(FP+FN), used in dataDL calculation
* @param predicted the default classification if ruleset covers null
* @return the combined class
*/
public double combinedDL(double expFPRate, double predicted) {
double rt = 0;
if (getRulesetSize() > 0) {
double[] stats = (double[]) m_SimpleStats.lastElement();
for (int j = getRulesetSize() - 2; j >= 0; j--) {
stats[0] += getSimpleStats(j)[0];
stats[2] += getSimpleStats(j)[2];
stats[4] += getSimpleStats(j)[4];
}
rt += dataDL(expFPRate, stats[0], stats[1], stats[4], stats[5]); // Data DL
} else { // Null coverage ruleset
double fn = 0.0;
for (int j = 0; j < m_Data.numInstances(); j++)
if ((int) m_Data.instance(j).classValue() == (int) predicted)
fn += m_Data.instance(j).weight();
rt += dataDL(expFPRate, 0.0, m_Data.sumOfWeights(), 0.0, fn);
}
for (int i = 0; i < getRulesetSize(); i++) // Theory DL
rt += theoryDL(i);
return rt;
}
/**
* Generates the classifier.
*
* @param instances set of instances serving as training data
* @throws Exception if the classifier has not been generated successfully
*/
public void buildClassifier(Instances instances) throws Exception {
// can classifier handle the data?
getCapabilities().testWithFail(instances);
// remove instances with missing class
instances = new Instances(instances);
instances.deleteWithMissingClass();
m_headerInfo = new Instances(instances, 0);
m_numClasses = instances.numClasses();
m_numAttributes = instances.numAttributes();
m_probOfWordGivenClass = new double[m_numClasses][];
/*
initialising the matrix of word counts
NOTE: Laplace estimator introduced in case a word that does not appear for a class in the
training set does so for the test set
*/
for (int c = 0; c < m_numClasses; c++) {
m_probOfWordGivenClass[c] = new double[m_numAttributes];
for (int att = 0; att < m_numAttributes; att++) {
m_probOfWordGivenClass[c][att] = 1;
}
}
// enumerate through the instances
Instance instance;
int classIndex;
double numOccurences;
double[] docsPerClass = new double[m_numClasses];
double[] wordsPerClass = new double[m_numClasses];
java.util.Enumeration enumInsts = instances.enumerateInstances();
while (enumInsts.hasMoreElements()) {
instance = (Instance) enumInsts.nextElement();
classIndex = (int) instance.value(instance.classIndex());
docsPerClass[classIndex] += instance.weight();
for (int a = 0; a < instance.numValues(); a++)
if (instance.index(a) != instance.classIndex()) {
if (!instance.isMissing(a)) {
numOccurences = instance.valueSparse(a) * instance.weight();
if (numOccurences < 0)
throw new Exception("Numeric attribute values must all be greater or equal to zero.");
wordsPerClass[classIndex] += numOccurences;
m_probOfWordGivenClass[classIndex][instance.index(a)] += numOccurences;
}
}
}
/*
normalising probOfWordGivenClass values
and saving each value as the log of each value
*/
for (int c = 0; c < m_numClasses; c++)
for (int v = 0; v < m_numAttributes; v++)
m_probOfWordGivenClass[c][v] =
Math.log(m_probOfWordGivenClass[c][v] / (wordsPerClass[c] + m_numAttributes - 1));
/*
calculating Pr(H)
NOTE: Laplace estimator introduced in case a class does not get mentioned in the set of
training instances
*/
final double numDocs = instances.sumOfWeights() + m_numClasses;
m_probOfClass = new double[m_numClasses];
for (int h = 0; h < m_numClasses; h++)
m_probOfClass[h] = (double) (docsPerClass[h] + 1) / numDocs;
}