nd.java

代码是一个分类器的实现,其中使用了部分weka的源代码。可以将项目导入eclipse运行

/*
 *    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
 *
 */

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.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;

/**
 <!-- 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>
 * &#64;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}
 * }
 * 
 * &#64;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 &lt;num&gt;
 *  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 &lt;pruning confidence&gt;
 *  Set confidence threshold for pruning.
 *  (default 0.25)</pre>
 * 
 * <pre> -M &lt;minimum number of instances&gt;
 *  Set minimum number of instances per leaf.
 *  (default 2)</pre>
 * 
 * <pre> -R
 *  Use reduced error pruning.</pre>
 * 
 * <pre> -N &lt;number of folds&gt;
 *  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 &lt;seed&gt;
 *  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 {
    
    /** 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();