/*
* 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.
*/
/*
* ND.java
* Copyright (C) 2003-2005 University of Waikato, Hamilton, New Zealand
*
*/
package weka.classifiers.meta.nestedDichotomies;
import weka.classifiers.Classifier;
import weka.classifiers.RandomizableSingleClassifierEnhancer;
import weka.classifiers.meta.FilteredClassifier;
import weka.classifiers.rules.ZeroR;
import weka.core.Capabilities;
import weka.core.FastVector;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.RevisionHandler;
import weka.core.RevisionUtils;
import weka.core.TechnicalInformation;
import weka.core.TechnicalInformationHandler;
import weka.core.Capabilities.Capability;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.MakeIndicator;
import weka.filters.unsupervised.instance.RemoveWithValues;
import java.io.Serializable;
import java.util.Hashtable;
import java.util.Random;
import weka.classifiers.AbstractClassifier;
/**
<!-- globalinfo-start -->
* A meta classifier for handling multi-class datasets with 2-class classifiers by building a random tree structure.<br/>
* <br/>
* For more info, check<br/>
* <br/>
* Lin Dong, Eibe Frank, Stefan Kramer: Ensembles of Balanced Nested Dichotomies for Multi-class Problems. In: PKDD, 84-95, 2005.<br/>
* <br/>
* Eibe Frank, Stefan Kramer: Ensembles of nested dichotomies for multi-class problems. In: Twenty-first International Conference on Machine Learning, 2004.
* <p/>
<!-- globalinfo-end -->
*
<!-- technical-bibtex-start -->
* BibTeX:
* <pre>
* @inproceedings{Dong2005,
* author = {Lin Dong and Eibe Frank and Stefan Kramer},
* booktitle = {PKDD},
* pages = {84-95},
* publisher = {Springer},
* title = {Ensembles of Balanced Nested Dichotomies for Multi-class Problems},
* year = {2005}
* }
*
* @inproceedings{Frank2004,
* author = {Eibe Frank and Stefan Kramer},
* booktitle = {Twenty-first International Conference on Machine Learning},
* publisher = {ACM},
* title = {Ensembles of nested dichotomies for multi-class problems},
* year = {2004}
* }
* </pre>
* <p/>
<!-- technical-bibtex-end -->
*
<!-- options-start -->
* Valid options are: <p/>
*
* <pre> -S <num>
* Random number seed.
* (default 1)</pre>
*
* <pre> -D
* If set, classifier is run in debug mode and
* may output additional info to the console</pre>
*
* <pre> -W
* Full name of base classifier.
* (default: weka.classifiers.trees.J48)</pre>
*
* <pre>
* Options specific to classifier weka.classifiers.trees.J48:
* </pre>
*
* <pre> -U
* Use unpruned tree.</pre>
*
* <pre> -C <pruning confidence>
* Set confidence threshold for pruning.
* (default 0.25)</pre>
*
* <pre> -M <minimum number of instances>
* Set minimum number of instances per leaf.
* (default 2)</pre>
*
* <pre> -R
* Use reduced error pruning.</pre>
*
* <pre> -N <number of folds>
* Set number of folds for reduced error
* pruning. One fold is used as pruning set.
* (default 3)</pre>
*
* <pre> -B
* Use binary splits only.</pre>
*
* <pre> -S
* Don't perform subtree raising.</pre>
*
* <pre> -L
* Do not clean up after the tree has been built.</pre>
*
* <pre> -A
* Laplace smoothing for predicted probabilities.</pre>
*
* <pre> -Q <seed>
* Seed for random data shuffling (default 1).</pre>
*
<!-- options-end -->
*
* @author Eibe Frank
* @author Lin Dong
*/
public class ND
extends RandomizableSingleClassifierEnhancer
implements TechnicalInformationHandler {
/** for serialization */
static final long serialVersionUID = -6355893369855683820L;
/**
* a node class
*/
protected class NDTree
implements Serializable, RevisionHandler {
/** for serialization */
private static final long serialVersionUID = 4284655952754474880L;
/** The indices associated with this node */
protected FastVector m_indices = null;
/** The parent */
protected NDTree m_parent = null;
/** The left successor */
protected NDTree m_left = null;
/** The right successor */
protected NDTree m_right = null;
/**
* Constructor.
*/
protected NDTree() {
m_indices = new FastVector(1);
m_indices.addElement(new Integer(Integer.MAX_VALUE));
}
/**
* Locates the node with the given index (depth-first traversal).
*/
protected NDTree locateNode(int nodeIndex, int[] currentIndex) {
if (nodeIndex == currentIndex[0]) {
return this;
} else if (m_left == null) {
return null;
} else {
currentIndex[0]++;
NDTree leftresult = m_left.locateNode(nodeIndex, currentIndex);
if (leftresult != null) {
return leftresult;
} else {
currentIndex[0]++;
return m_right.locateNode(nodeIndex, currentIndex);
}
}
}
/**
* Inserts a class index into the tree.
*
* @param classIndex the class index to insert
*/
protected void insertClassIndex(int classIndex) {
// Create new nodes
NDTree right = new NDTree();
if (m_left != null) {
m_right.m_parent = right;
m_left.m_parent = right;
right.m_right = m_right;
right.m_left = m_left;
}
m_right = right;
m_right.m_indices = (FastVector)m_indices.copy();
m_right.m_parent = this;
m_left = new NDTree();
m_left.insertClassIndexAtNode(classIndex);
m_left.m_parent = this;
// Propagate class Index
propagateClassIndex(classIndex);
}
/**
* Propagates class index to the root.
*
* @param classIndex the index to propagate to the root
*/
protected void propagateClassIndex(int classIndex) {
insertClassIndexAtNode(classIndex);
if (m_parent != null) {
m_parent.propagateClassIndex(classIndex);
}
}
/**
* Inserts the class index at a given node.
*
* @param classIndex the classIndex to insert
*/
protected void insertClassIndexAtNode(int classIndex) {
int i = 0;
while (classIndex > ((Integer)m_indices.elementAt(i)).intValue()) {
i++;
}
m_indices.insertElementAt(new Integer(classIndex), i);
}
/**
* Gets the indices in an array of ints.
*
* @return the indices
*/
protected int[] getIndices() {
int[] ints = new int[m_indices.size() - 1];
for (int i = 0; i < m_indices.size() - 1; i++) {
ints[i] = ((Integer)m_indices.elementAt(i)).intValue();
}
return ints;
}
/**
* Checks whether an index is in the array.
*
* @param index the index to check
* @return true of the index is in the array
*/
protected boolean contains(int index) {
for (int i = 0; i < m_indices.size() - 1; i++) {
if (index == ((Integer)m_indices.elementAt(i)).intValue()) {
return true;
}
}
return false;
}
/**
* Returns the list of indices as a string.
*
* @return the indices as string
*/
protected String getString() {
StringBuffer string = new StringBuffer();
for (int i = 0; i < m_indices.size() - 1; i++) {
if (i > 0) {
string.append(',');
}
string.append(((Integer)m_indices.elementAt(i)).intValue() + 1);
}
return string.toString();
}
/**
* Unifies tree for improve hashing.
*/
protected void unifyTree() {
if (m_left != null) {
if (((Integer)m_left.m_indices.elementAt(0)).intValue() >
((Integer)m_right.m_indices.elementAt(0)).intValue()) {
NDTree temp = m_left;
m_left = m_right;
m_right = temp;
}
m_left.unifyTree();
m_right.unifyTree();
}
}
/**
* Returns a description of the tree rooted at this node.
*
* @param text the buffer to add the node to
* @param id the node id
* @param level the level of the tree
*/
protected void toString(StringBuffer text, int[] id, int level) {
for (int i = 0; i < level; i++) {
text.append(" | ");
}
text.append(id[0] + ": " + getString() + "\n");
if (m_left != null) {
id[0]++;
m_left.toString(text, id, level + 1);
id[0]++;
m_right.toString(text, id, level + 1);
}
}
/**
* Returns the revision string.
*
* @return the revision
*/
public String getRevision() {
return RevisionUtils.extract("$Revision: 1.9 $");
}
}
/** The tree of classes */
protected NDTree m_ndtree = null;
/** The hashtable containing all the classifiers */
protected Hashtable m_classifiers = null;
/** Is Hashtable given from END? */
protected boolean m_hashtablegiven = false;
/**
* Constructor.
*/
public ND() {
m_Classifier = new weka.classifiers.trees.J48();
}
/**
* String describing default classifier.
*
* @return the default classifier classname
*/
protected String defaultClassifierString() {
return "weka.classifiers.trees.J48";
}
/**
* 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
*/
public TechnicalInformation getTechnicalInformation() {
TechnicalInformation result;
TechnicalInformation additional;
result = new TechnicalInformation(Type.INPROCEEDINGS);
result.setValue(Field.AUTHOR, "Lin Dong and Eibe Frank and Stefan Kramer");
result.setValue(Field.TITLE, "Ensembles of Balanced Nested Dichotomies for Multi-class Problems");
result.setValue(Field.BOOKTITLE, "PKDD");
result.setValue(Field.YEAR, "2005");
result.setValue(Field.PAGES, "84-95");
result.setValue(Field.PUBLISHER, "Springer");
additional = result.add(Type.INPROCEEDINGS);
additional.setValue(Field.AUTHOR, "Eibe Frank and Stefan Kramer");
additional.setValue(Field.TITLE, "Ensembles of nested dichotomies for multi-class problems");
additional.setValue(Field.BOOKTITLE, "Twenty-first International Conference on Machine Learning");
additional.setValue(Field.YEAR, "2004");
additional.setValue(Field.PUBLISHER, "ACM");
return result;
}
/**
* Set hashtable from END.
*
* @param table the hashtable to use
*/
public void setHashtable(Hashtable table) {
m_hashtablegiven = true;
m_classifiers = table;
}
/**
* Returns default capabilities of the classifier.
*
* @return the capabilities of this classifier
*/
public Capabilities getCapabilities() {
Capabilities result = super.getCapabilities();
// class
result.disableAllClasses();
result.enable(Capability.NOMINAL_CLASS);
result.enable(Capability.MISSING_CLASS_VALUES);
// instances
result.setMinimumNumberInstances(1);
return result;
}
/**
* Builds the classifier.
*
* @param data the data to train the classifier with
* @throws Exception if anything goes wrong
*/
public void buildClassifier(Instances data) throws Exception {
// can classifier handle the data?
getCapabilities().testWithFail(data);
// remove instances with missing class
data = new Instances(data);
data.deleteWithMissingClass();
Random random = data.getRandomNumberGenerator(m_Seed);
if (!m_hashtablegiven) {
m_classifiers = new Hashtable();
}
// Generate random class hierarchy
int[] indices = new int[data.numClasses()];
for (int i = 0; i < indices.length; i++) {
indices[i] = i;
}
// Randomize list of class indices
for (int i = indices.length - 1; i > 0; i--) {
int help = indices[i];
int index = random.nextInt(i + 1);
indices[i] = indices[index];
indices[index] = help;
}
// Insert random class index at randomly chosen node
m_ndtree = new NDTree();
m_ndtree.insertClassIndexAtNode(indices[0]);
for (int i = 1; i < indices.length; i++) {
int nodeIndex = random.nextInt(2 * i - 1);
NDTree node = m_ndtree.locateNode(nodeIndex, new int[1]);
node.insertClassIndex(indices[i]);
}
m_ndtree.unifyTree();
// Build classifiers
buildClassifierForNode(m_ndtree, data);
}
/**
* Builds the classifier for one node.
*
* @param node the node to build the classifier for
* @param data the data to work with
* @throws Exception if anything goes wrong
*/
public void buildClassifierForNode(NDTree node, Instances data) throws Exception {
// Are we at a leaf node ?
if (node.m_left != null) {
// Create classifier
MakeIndicator filter = new MakeIndicator();
filter.setAttributeIndex("" + (data.classIndex() + 1));
filter.setValueIndices(node.m_right.getString());
filter.setNumeric(false);
filter.setInputFormat(data);
FilteredClassifier classifier = new FilteredClassifier();
if (data.numInstances() > 0) {
classifier.setClassifier(AbstractClassifier.makeCopies(m_Classifier, 1)[0]);
} else {
classifier.setClassifier(new ZeroR());
}
classifier.setFilter(filter);
if (!m_classifiers.containsKey(node.m_left.getString() + "|" + node.m_right.getString())) {
classifier.buildClassifier(data);
m_classifiers.put(node.m_left.getString() + "|" + node.m_right.getString(), classifier);
} else {
classifier=(FilteredClassifier)m_classifiers.get(node.m_left.getString() + "|" +
node.m_right.getString());
}
// Generate successors
if (node.m_left.m_left != null) {
RemoveWithValues rwv = new RemoveWithValues();
rwv.setInvertSelection(true);
rwv.setNominalIndices(node.m_left.getString());
rwv.setAttributeIndex("" + (data.classIndex() + 1));
rwv.setInputFormat(data);
Instances firstSubset = Filter.useFilter(data, rwv);
buildClassifierForNode(node.m_left, firstSubset);
}
if (node.m_right.m_left != null) {
RemoveWithValues rwv = new RemoveWithValues();
rwv.setInvertSelection(true);
rwv.setNominalIndices(node.m_right.getString());
rwv.setAttributeIndex("" + (data.classIndex() + 1));
rwv.setInputFormat(data);
Instances secondSubset = Filter.useFilter(data, rwv);
buildClassifierForNode(node.m_right, secondSubset);
}
}
}
/**
* Predicts the class distribution for a given instance
*
* @param inst the (multi-class) instance to be classified
* @return the class distribution
* @throws Exception if computing fails
*/
public double[] distributionForInstance(Instance inst) throws Exception {
return distributionForInstance(inst, m_ndtree);
}
/**
* Predicts the class distribution for a given instance
*
* @param inst the (multi-class) instance to be classified
* @param node the node to do get the distribution for
* @return the class distribution
* @throws Exception if computing fails
*/
protected double[] distributionForInstance(Instance inst, NDTree node) throws Exception {
double[] newDist = new double[inst.numClasses()];
if (node.m_left == null) {
newDist[node.getIndices()[0]] = 1.0;
return newDist;
} else {
Classifier classifier = (Classifier)m_classifiers.get(node.m_left.getString() + "|" +
node.m_right.getString());
double[] leftDist = distributionForInstance(inst, node.m_left);
double[] rightDist = distributionForInstance(inst, node.m_right);
double[] dist = classifier.distributionForInstance(inst);
for (int i = 0; i < inst.numClasses(); i++) {
if (node.m_right.contains(i)) {
newDist[i] = dist[1] * rightDist[i];
} else {
newDist[i] = dist[0] * leftDist[i];
}
}
return newDist;
}
}
/**
* Outputs the classifier as a string.
*
* @return a string representation of the classifier
*/
public String toString() {
if (m_classifiers == null) {
return "ND: No model built yet.";
}
StringBuffer text = new StringBuffer();
text.append("ND\n\n");
m_ndtree.toString(text, new int[1], 0);
return text.toString();
}
/**
* @return a description of the classifier suitable for
* displaying in the explorer/experimenter gui
*/
public String globalInfo() {
return
"A meta classifier for handling multi-class datasets with 2-class "
+ "classifiers by building a random tree structure.\n\n"
+ "For more info, check\n\n"
+ getTechnicalInformation().toString();
}
/**
* Returns the revision string.
*
* @return the revision
*/
public String getRevision() {
return RevisionUtils.extract("$Revision: 1.9 $");
}
/**
* Main method for testing this class.
*
* @param argv the options
*/
public static void main(String [] argv) {
runClassifier(new ND(), argv);
}
}