j48.java

wekaUT是 university texas austin 开发的基于weka的半指导学习(semi supervised learning)的分类器

/*
 *    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.
 */

/*
 *    J48.java
 *    Copyright (C) 1999 Eibe Frank
 *
 */

package weka.classifiers.trees.j48;

import java.util.*;
import weka.core.*;
import weka.classifiers.*;

/**
 * Class for generating an unpruned or a pruned C4.5 decision tree.
 * For more information, see<p>
 *
 * Ross Quinlan (1993). <i>C4.5: Programs for Machine Learning</i>, 
 * Morgan Kaufmann Publishers, San Mateo, CA. </p>
 *
 * Valid options are: <p>
 *
 * -U <br>
 * Use unpruned tree.<p>
 *
 * -C confidence <br>
 * Set confidence threshold for pruning. (Default: 0.25) <p>
 *
 * -M number <br>
 * Set minimum number of instances per leaf. (Default: 2) <p>
 *
 * -R <br>
 * Use reduced error pruning. No subtree raising is performed. <p>
 *
 * -N number <br>
 * Set number of folds for reduced error pruning. One fold is
 * used as the pruning set. (Default: 3) <p>
 *
 * -B <br>
 * Use binary splits for nominal attributes. <p>
 *
 * -S <br>
 * Don't perform subtree raising. <p>
 *
 * -L <br>
 * Do not clean up after the tree has been built. <p>
 *
 * -A <br>
 * If set, Laplace smoothing is used for predicted probabilites. <p>
 *
 *
 * @author Eibe Frank ([e-mail])
 * @version $Revision: 1.1.1.1 $
 */
public class J48 extends DistributionClassifier implements OptionHandler, 
  Drawable, Matchable, Sourcable, WeightedInstancesHandler, Summarizable,
  AdditionalMeasureProducer {

  // To maintain the same version number after adding m_ClassAttribute
  static final long serialVersionUID = -217733168393644444L;

  /** The decision tree */
  private ClassifierTree m_root;
  
  /** Unpruned tree? */
  private boolean m_unpruned = false;

  /** Confidence level */
  private float m_CF = 0.25f;

  /** Minimum number of instances */
  private int m_minNumObj = 2;

  /** Determines whether probabilities are smoothed using
      Laplace correction when predictions are generated */
  private boolean m_useLaplace = false;

  /** Use reduced error pruning? */
  private boolean m_reducedErrorPruning = false;

  /** Number of folds for reduced error pruning. */
  private int m_numFolds = 3;

  /** Binary splits on nominal attributes? */
  private boolean m_binarySplits = false;

  /** Subtree raising to be performed? */
  private boolean m_subtreeRaising = true;

  /** Cleanup after the tree has been built. */
  boolean m_noCleanup = false;
  
  /**
   * Generates the classifier.
   *
   * @exception Exception if classifier can't be built successfully
   */
  public void buildClassifier(Instances instances) 
       throws Exception {

    ModelSelection modSelection;	 

    if (m_binarySplits)
      modSelection = new BinC45ModelSelection(m_minNumObj, instances);
    else
      modSelection = new C45ModelSelection(m_minNumObj, instances);
    if (!m_reducedErrorPruning)
      m_root = new C45PruneableClassifierTree(modSelection, !m_unpruned, m_CF,
					    m_subtreeRaising, !m_noCleanup);
    else
      m_root = new PruneableClassifierTree(modSelection, !m_unpruned, m_numFolds,
					   !m_noCleanup);
    m_root.buildClassifier(instances);
    if (m_binarySplits) {
      ((BinC45ModelSelection)modSelection).cleanup();
    } else {
      ((C45ModelSelection)modSelection).cleanup();
    }
  }

  /**
   * Classifies an instance.
   *
   * @exception Exception if instance can't be classified successfully
   */
  public double classifyInstance(Instance instance) throws Exception {

    return m_root.classifyInstance(instance);
  }

  /** 
   * Returns class probabilities for an instance.
   *
   * @exception Exception if distribution can't be computed successfully
   */
  public final double [] distributionForInstance(Instance instance) 
       throws Exception {

    return m_root.distributionForInstance(instance, m_useLaplace);
  }

  /**
   * Returns graph describing the tree.
   *
   * @exception Exception if graph can't be computed
   */
  public String graph() throws Exception {

    return m_root.graph();
  }

  /**
   * Returns tree in prefix order.
   *
   * @exception Exception if something goes wrong
   */
  public String prefix() throws Exception {
    
    return m_root.prefix();
  }


  /**
   * Returns tree as an if-then statement.
   *
   * @return the tree as a Java if-then type statement
   * @exception Exception if something goes wrong
   */
  public String toSource(String className) throws Exception {

    StringBuffer [] source = m_root.toSource(className);
    return 
    "class " + className + " {\n\n"
    +"  public static double classify(Object [] i)\n"
    +"    throws Exception {\n\n"
    +"    double p = Double.NaN;\n"
    + source[0]  // Assignment code
    +"    return p;\n"
    +"  }\n"
    + source[1]  // Support code
    +"}\n";
  }

  /**
   * Returns an enumeration describing the available options.
   *
   * Valid options are: <p>
   *
   * -U <br>
   * Use unpruned tree.<p>
   *
   * -C confidence <br>
   * Set confidence threshold for pruning. (Default: 0.25) <p>
   *
   * -M number <br>
   * Set minimum number of instances per leaf. (Default: 2) <p>
   *
   * -R <br>
   * Use reduced error pruning. No subtree raising is performed. <p>
   *
   * -N number <br>
   * Set number of folds for reduced error pruning. One fold is
   * used as the pruning set. (Default: 3) <p>
   *
   * -B <br>
   * Use binary splits for nominal attributes. <p>
   *
   * -S <br>
   * Don't perform subtree raising. <p>
   *
   * -L <br>
   * Do not clean up after the tree has been built.
   *
   * -A <br>
   * If set, Laplace smoothing is used for predicted probabilites. <p>
   *
   * @return an enumeration of all the available options.
   */
  public Enumeration listOptions() {

    Vector newVector = new Vector(9);

    newVector.
	addElement(new Option("\tUse unpruned tree.",
			      "U", 0, "-U"));
    newVector.
	addElement(new Option("\tSet confidence threshold for pruning.\n" +
			      "\t(default 0.25)",
			      "C", 1, "-C <pruning confidence>"));
    newVector.
	addElement(new Option("\tSet minimum number of instances per leaf.\n" +
			      "\t(default 2)",
			      "M", 1, "-M <minimum number of instances>"));
    newVector.
	addElement(new Option("\tUse reduced error pruning.",
			      "R", 0, "-R"));
    newVector.
	addElement(new Option("\tSet number of folds for reduced error\n" +
			      "\tpruning. One fold is used as pruning set.\n" +
			      "\t(default 3)",
			      "N", 1, "-N <number of folds>"));
    newVector.
	addElement(new Option("\tUse binary splits only.",
			      "B", 0, "-B"));
    newVector.
        addElement(new Option("\tDon't perform subtree raising.",
			      "S", 0, "-S"));
    newVector.
        addElement(new Option("\tDo not clean up after the tree has been built.",
			      "L", 0, "-L"));
   newVector.
        addElement(new Option("\tLaplace smoothing for predicted probabilities.",
			      "A", 0, "-A"));

    return newVector.elements();
  }

  /**
   * Parses a given list of options.
   *
   * @param options the list of options as an array of strings
   * @exception Exception if an option is not supported
   */
  public void setOptions(String[] options) throws Exception {
    
    // Other options
    String minNumString = Utils.getOption('M', options);
    if (minNumString.length() != 0) {
      m_minNumObj = Integer.parseInt(minNumString);
    } else {
      m_minNumObj = 2;
    }
    m_binarySplits = Utils.getFlag('B', options);
    m_useLaplace = Utils.getFlag('A', options);

    // Pruning options
    m_unpruned = Utils.getFlag('U', options);
    m_subtreeRaising = !Utils.getFlag('S', options);
    m_noCleanup = Utils.getFlag('L', options);
    if ((m_unpruned) && (!m_subtreeRaising)) {
      throw new Exception("Subtree raising doesn't need to be unset for unpruned tree!");
    }
    m_reducedErrorPruning = Utils.getFlag('R', options);
    if ((m_unpruned) && (m_reducedErrorPruning)) {
      throw new Exception("Unpruned tree and reduced error pruning can't be selected " +
			  "simultaneously!");
    }
    String confidenceString = Utils.getOption('C', options);
    if (confidenceString.length() != 0) {
      if (m_reducedErrorPruning) {
	throw new Exception("Setting the confidence doesn't make sense " +
			    "for reduced error pruning.");
      } else if (m_unpruned) {
	throw new Exception("Doesn't make sense to change confidence for unpruned "
			    +"tree!");
      } else {
	m_CF = (new Float(confidenceString)).floatValue();
	if ((m_CF <= 0) || (m_CF >= 1)) {
	  throw new Exception("Confidence has to be greater than zero and smaller " +
			      "than one!");
	}
      }
    } else {
      m_CF = 0.25f;
    }
    String numFoldsString = Utils.getOption('N', options);
    if (numFoldsString.length() != 0) {
      if (!m_reducedErrorPruning) {
	throw new Exception("Setting the number of folds" +
			    " doesn't make sense if" +
			    " reduced error pruning is not selected.");
      } else {
	m_numFolds = Integer.parseInt(numFoldsString);
      }
    } else {
      m_numFolds = 3;