adaboostm1.java

:<<数据挖掘--实用机器学习技术及java实现>>一书的配套源程序

/*
 *    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.
 */

/*
 *    AdaBoostM1.java
 *    Copyright (C) 1999 Eibe Frank,Len Trigg
 *
 */

package weka.classifiers;

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

/**
 * Class for boosting a classifier using Freund & Schapire's Adaboost 
 * M1 method. For more information, see<p>
 *
 * Yoav Freund and Robert E. Schapire
 * (1996). <i>Experiments with a new boosting algorithm</i>.  Proc
 * International Conference on Machine Learning, pages 148-156, Morgan
 * Kaufmann, San Francisco.<p>
 *
 * Valid options are:<p>
 *
 * -D <br>
 * Turn on debugging output.<p>
 *
 * -W classname <br>
 * Specify the full class name of a classifier as the basis for 
 * boosting (required).<p>
 *
 * -I num <br>
 * Set the number of boost iterations (default 10). <p>
 *
 * -P num <br>
 * Set the percentage of weight mass used to build classifiers
 * (default 100). <p>
 *
 * -Q <br>
 * Use resampling instead of reweighting.<p>
 *
 * -S seed <br>
 * Random number seed for resampling (default 1). <p>
 *
 * Options after -- are passed to the designated classifier.<p>
 *
 * @author Eibe Frank ([e-mail])
 * @author Len Trigg ([e-mail])
 * @version $Revision: 1.12 $ 
 */
public class AdaBoostM1 extends DistributionClassifier 
  implements OptionHandler, WeightedInstancesHandler, Sourcable {

  /** Max num iterations tried to find classifier with non-zero error. */ 
  private static int MAX_NUM_RESAMPLING_ITERATIONS = 10;

  /** The model base classifier to use */
  protected Classifier m_Classifier = new weka.classifiers.ZeroR();
  
  /** Array for storing the generated base classifiers. */
  protected Classifier [] m_Classifiers;
  
  /** Array for storing the weights for the votes. */
  protected double [] m_Betas;

  /** The maximum number of boost iterations */
  protected int m_MaxIterations = 10;

  /** The number of successfully generated base classifiers. */
  protected int m_NumIterations;

  /** Weight Threshold. The percentage of weight mass used in training */
  protected int m_WeightThreshold = 100;

  /** Debugging mode, gives extra output if true */
  protected boolean m_Debug;

  /** Use boosting with reweighting? */
  protected boolean m_UseResampling;

  /** Seed for boosting with resampling. */
  protected int m_Seed = 1;

  /** The number of classes */
  protected int m_NumClasses;

  /**
   * Select only instances with weights that contribute to 
   * the specified quantile of the weight distribution
   *
   * @param data the input instances
   * @param quantile the specified quantile eg 0.9 to select 
   * 90% of the weight mass
   * @return the selected instances
   */
  protected Instances selectWeightQuantile(Instances data, double quantile) { 

    int numInstances = data.numInstances();
    Instances trainData = new Instances(data, numInstances);
    double [] weights = new double [numInstances];

    double sumOfWeights = 0;
    for(int i = 0; i < numInstances; i++) {
      weights[i] = data.instance(i).weight();
      sumOfWeights += weights[i];
    }
    double weightMassToSelect = sumOfWeights * quantile;
    int [] sortedIndices = Utils.sort(weights);

    // Select the instances
    sumOfWeights = 0;
    for(int i = numInstances - 1; i >= 0; i--) {
      Instance instance = (Instance)data.instance(sortedIndices[i]).copy();
      trainData.add(instance);
      sumOfWeights += weights[sortedIndices[i]];
      if ((sumOfWeights > weightMassToSelect) && 
	  (i > 0) && 
	  (weights[sortedIndices[i]] != weights[sortedIndices[i - 1]])) {
	break;
      }
    }
    if (m_Debug) {
      System.err.println("Selected " + trainData.numInstances()
			 + " out of " + numInstances);
    }
    return trainData;
  }

  /**
   * Returns an enumeration describing the available options
   *
   * @return an enumeration of all the available options
   */
  public Enumeration listOptions() {

    Vector newVector = new Vector(6);

    newVector.addElement(new Option(
	      "\tTurn on debugging output.",
	      "D", 0, "-D"));
    newVector.addElement(new Option(
	      "\tMaximum number of boost iterations.\n"
	      +"\t(default 10)",
	      "I", 1, "-I <num>"));
    newVector.addElement(new Option(
	      "\tPercentage of weight mass to base training on.\n"
	      +"\t(default 100, reduce to around 90 speed up)",
	      "P", 1, "-P <num>"));
    newVector.addElement(new Option(
	      "\tFull name of classifier to boost.\n"
	      +"\teg: weka.classifiers.NaiveBayes",
	      "W", 1, "-W <class name>"));
    newVector.addElement(new Option(
	      "\tUse resampling for boosting.",
	      "Q", 0, "-Q"));
    newVector.addElement(new Option(
	      "\tSeed for resampling. (Default 1)",
	      "S", 1, "-S <num>"));
    

    if ((m_Classifier != null) &&
	(m_Classifier instanceof OptionHandler)) {
      newVector.addElement(new Option(
	     "",
	     "", 0, "\nOptions specific to classifier "
	     + m_Classifier.getClass().getName() + ":"));
      Enumeration enum = ((OptionHandler)m_Classifier).listOptions();
      while (enum.hasMoreElements()) {
	newVector.addElement(enum.nextElement());
      }
    }
    return newVector.elements();
  }


  /**
   * Parses a given list of options. Valid options are:<p>
   *
   * -D <br>
   * Turn on debugging output.<p>
   *
   * -W classname <br>
   * Specify the full class name of a classifier as the basis for 
   * boosting (required).<p>
   *
   * -I num <br>
   * Set the number of boost iterations (default 10). <p>
   *
   * -P num <br>
   * Set the percentage of weight mass used to build classifiers
   * (default 100). <p>
   *
   * -Q <br>
   * Use resampling instead of reweighting.<p>
   *
   * -S seed <br>
   * Random number seed for resampling (default 1).<p>
   *
   * Options after -- are passed to the designated classifier.<p>
   *
   * @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 {
    
    setDebug(Utils.getFlag('D', options));
    
    String boostIterations = Utils.getOption('I', options);
    if (boostIterations.length() != 0) {
      setMaxIterations(Integer.parseInt(boostIterations));
    } else {
      setMaxIterations(10);
    }

    String thresholdString = Utils.getOption('P', options);
    if (thresholdString.length() != 0) {
      setWeightThreshold(Integer.parseInt(thresholdString));
    } else {
      setWeightThreshold(100);
    }
      
    setUseResampling(Utils.getFlag('Q', options));
    if (m_UseResampling && (thresholdString.length() != 0)) {
      throw new Exception("Weight pruning with resampling"+
			  "not allowed.");
    }

    String seedString = Utils.getOption('S', options);
    if (seedString.length() != 0) {
      setSeed(Integer.parseInt(seedString));
    } else {
      setSeed(1);
    }

    String classifierName = Utils.getOption('W', options);
    if (classifierName.length() == 0) {
      throw new Exception("A classifier must be specified with"
			  + " the -W option.");
    }
    setClassifier(Classifier.forName(classifierName,
				     Utils.partitionOptions(options)));
  }

  /**
   * Gets the current settings of the Classifier.
   *
   * @return an array of strings suitable for passing to setOptions
   */
  public String [] getOptions() {

    String [] classifierOptions = new String [0];
    if ((m_Classifier != null) && 
	(m_Classifier instanceof OptionHandler)) {
      classifierOptions = ((OptionHandler)m_Classifier).getOptions();
    }

    String [] options = new String [classifierOptions.length + 10];
    int current = 0;
    if (getDebug()) {
      options[current++] = "-D";
    }
    if (getUseResampling()) {
      options[current++] = "-Q";
    } else {
      options[current++] = "-P"; 
      options[current++] = "" + getWeightThreshold();
    }
    options[current++] = "-I"; options[current++] = "" + getMaxIterations();
    options[current++] = "-S"; options[current++] = "" + getSeed();

    if (getClassifier() != null) {
      options[current++] = "-W";
      options[current++] = getClassifier().getClass().getName();
    }
    options[current++] = "--";

    System.arraycopy(classifierOptions, 0, options, current, 
		     classifierOptions.length);
    current += classifierOptions.length;
    while (current < options.length) {
      options[current++] = "";
    }
    return options;
  }

  /**
   * Set the classifier for boosting. 
   *
   * @param newClassifier the Classifier to use.
   */
  public void setClassifier(Classifier newClassifier) {

    m_Classifier = newClassifier;
  }

  /**
   * Get the classifier used as the classifier
   *
   * @return the classifier used as the classifier
   */
  public Classifier getClassifier() {

    return m_Classifier;
  }


  /**
   * Set the maximum number of boost iterations
   */
  public void setMaxIterations(int maxIterations) {

    m_MaxIterations = maxIterations;
  }

  /**
   * Get the maximum number of boost iterations
   *
   * @return the maximum number of boost iterations
   */
  public int getMaxIterations() {

    return m_MaxIterations;
  }

  /**
   * Set weight threshold
   *
   * @param thresholding the percentage of weight mass used for training
   */
  public void setWeightThreshold(int threshold) {

    m_WeightThreshold = threshold;
  }

  /**
   * Get the degree of weight thresholding
   *
   * @return the percentage of weight mass used for training
   */
  public int getWeightThreshold() {

    return m_WeightThreshold;
  }

  /**
   * Set seed for resampling.
   *
   * @param seed the seed for resampling
   */
  public void setSeed(int seed) {

    m_Seed = seed;
  }

  /**
   * Get seed for resampling.
   *
   * @return the seed for resampling
   */
  public int getSeed() {

    return m_Seed;
  }

  /**