logitboost.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.
 */

/*
 *    LogitBoost.java
 *    Copyright (C) 1999, 2002 Len Trigg, Eibe Frank
 *
 */

package weka.classifiers.meta;

import weka.classifiers.Classifier;
import weka.classifiers.Evaluation;
import weka.classifiers.RandomizableIteratedSingleClassifierEnhancer;
import weka.classifiers.Sourcable;
import weka.core.Attribute;
import weka.core.Capabilities;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.Option;
import weka.core.TechnicalInformation;
import weka.core.TechnicalInformationHandler;
import weka.core.Utils;
import weka.core.WeightedInstancesHandler;
import weka.core.Capabilities.Capability;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;

import java.util.Enumeration;
import java.util.Random;
import java.util.Vector;

/**
 <!-- globalinfo-start -->
 * Class for performing additive logistic regression. <br/>
 * This class performs classification using a regression scheme as the base learner, and can handle multi-class problems.  For more information, see<br/>
 * <br/>
 * J. Friedman, T. Hastie, R. Tibshirani (1998). Additive Logistic Regression: a Statistical View of Boosting. Stanford University.<br/>
 * <br/>
 * Can do efficient internal cross-validation to determine appropriate number of iterations.
 * <p/>
 <!-- globalinfo-end -->
 *
 <!-- technical-bibtex-start -->
 * BibTeX:
 * <pre>
 * &#64;techreport{Friedman1998,
 *    address = {Stanford University},
 *    author = {J. Friedman and T. Hastie and R. Tibshirani},
 *    title = {Additive Logistic Regression: a Statistical View of Boosting},
 *    year = {1998},
 *    PS = {http://www-stat.stanford.edu/~jhf/ftp/boost.ps}
 * }
 * </pre>
 * <p/>
 <!-- technical-bibtex-end -->
 *
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -Q
 *  Use resampling instead of reweighting for boosting.</pre>
 * 
 * <pre> -P &lt;percent&gt;
 *  Percentage of weight mass to base training on.
 *  (default 100, reduce to around 90 speed up)</pre>
 * 
 * <pre> -F &lt;num&gt;
 *  Number of folds for internal cross-validation.
 *  (default 0 -- no cross-validation)</pre>
 * 
 * <pre> -R &lt;num&gt;
 *  Number of runs for internal cross-validation.
 *  (default 1)</pre>
 * 
 * <pre> -L &lt;num&gt;
 *  Threshold on the improvement of the likelihood.
 *  (default -Double.MAX_VALUE)</pre>
 * 
 * <pre> -H &lt;num&gt;
 *  Shrinkage parameter.
 *  (default 1)</pre>
 * 
 * <pre> -S &lt;num&gt;
 *  Random number seed.
 *  (default 1)</pre>
 * 
 * <pre> -I &lt;num&gt;
 *  Number of iterations.
 *  (default 10)</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.DecisionStump)</pre>
 * 
 * <pre> 
 * Options specific to classifier weka.classifiers.trees.DecisionStump:
 * </pre>
 * 
 * <pre> -D
 *  If set, classifier is run in debug mode and
 *  may output additional info to the console</pre>
 * 
 <!-- options-end -->
 *
 * Options after -- are passed to the designated learner.<p>
 *
 * @author Len Trigg ([e-mail])
 * @author Eibe Frank ([e-mail])
 * @version $Revision: 1.37 $ 
 */
public class LogitBoost 
  extends RandomizableIteratedSingleClassifierEnhancer
  implements Sourcable, WeightedInstancesHandler, TechnicalInformationHandler {

  /** for serialization */
  static final long serialVersionUID = -3905660358715833753L;
  
  /** Array for storing the generated base classifiers. 
   Note: we are hiding the variable from IteratedSingleClassifierEnhancer*/
  protected Classifier [][] m_Classifiers;

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

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

  /** The number of folds for the internal cross-validation. */
  protected int m_NumFolds = 0;

  /** The number of runs for the internal cross-validation. */
  protected int m_NumRuns = 1;

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

  /** A threshold for responses (Friedman suggests between 2 and 4) */
  protected static final double Z_MAX = 3;

  /** Dummy dataset with a numeric class */
  protected Instances m_NumericClassData;

  /** The actual class attribute (for getting class names) */
  protected Attribute m_ClassAttribute;

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

  /** The threshold on the improvement of the likelihood */   
  protected double m_Precision = -Double.MAX_VALUE;

  /** The value of the shrinkage parameter */
  protected double m_Shrinkage = 1;

  /** The random number generator used */
  protected Random m_RandomInstance = null;

  /** The value by which the actual target value for the
      true class is offset. */
  protected double m_Offset = 0.0;
    
  /**
   * Returns a string describing classifier
   * @return a description suitable for
   * displaying in the explorer/experimenter gui
   */
  public String globalInfo() {

    return "Class for performing additive logistic regression. \n"
      + "This class performs classification using a regression scheme as the "
      + "base learner, and can handle multi-class problems.  For more "
      + "information, see\n\n"
      + getTechnicalInformation().toString() + "\n\n"
      + "Can do efficient internal cross-validation to determine "
      + "appropriate number of iterations.";
  }
    
  /**
   * Constructor.
   */
  public LogitBoost() {
    
    m_Classifier = new weka.classifiers.trees.DecisionStump();
  }

  /**
   * 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;
    
    result = new TechnicalInformation(Type.TECHREPORT);
    result.setValue(Field.AUTHOR, "J. Friedman and T. Hastie and R. Tibshirani");
    result.setValue(Field.YEAR, "1998");
    result.setValue(Field.TITLE, "Additive Logistic Regression: a Statistical View of Boosting");
    result.setValue(Field.ADDRESS, "Stanford University");
    result.setValue(Field.PS, "http://www-stat.stanford.edu/~jhf/ftp/boost.ps");
    
    return result;
  }

  /**
   * String describing default classifier.
   * 
   * @return the default classifier classname
   */
  protected String defaultClassifierString() {
    
    return "weka.classifiers.trees.DecisionStump";
  }

  /**
   * 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(
	      "\tUse resampling instead of reweighting for boosting.",
	      "Q", 0, "-Q"));
    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 <percent>"));
    newVector.addElement(new Option(
	      "\tNumber of folds for internal cross-validation.\n"
	      +"\t(default 0 -- no cross-validation)",
	      "F", 1, "-F <num>"));
    newVector.addElement(new Option(
	      "\tNumber of runs for internal cross-validation.\n"
	      +"\t(default 1)",
	      "R", 1, "-R <num>"));
    newVector.addElement(new Option(
	      "\tThreshold on the improvement of the likelihood.\n"
	      +"\t(default -Double.MAX_VALUE)",
	      "L", 1, "-L <num>"));
    newVector.addElement(new Option(
	      "\tShrinkage parameter.\n"
	      +"\t(default 1)",
	      "H", 1, "-H <num>"));

    Enumeration enu = super.listOptions();
    while (enu.hasMoreElements()) {
      newVector.addElement(enu.nextElement());
    }
    return newVector.elements();
  }


  /**
   * Parses a given list of options. <p/>
   * 
   <!-- options-start -->
   * Valid options are: <p/>
   * 
   * <pre> -Q
   *  Use resampling instead of reweighting for boosting.</pre>
   * 
   * <pre> -P &lt;percent&gt;
   *  Percentage of weight mass to base training on.
   *  (default 100, reduce to around 90 speed up)</pre>
   * 
   * <pre> -F &lt;num&gt;
   *  Number of folds for internal cross-validation.
   *  (default 0 -- no cross-validation)</pre>
   * 
   * <pre> -R &lt;num&gt;
   *  Number of runs for internal cross-validation.
   *  (default 1)</pre>
   * 
   * <pre> -L &lt;num&gt;
   *  Threshold on the improvement of the likelihood.
   *  (default -Double.MAX_VALUE)</pre>
   * 
   * <pre> -H &lt;num&gt;
   *  Shrinkage parameter.
   *  (default 1)</pre>
   * 
   * <pre> -S &lt;num&gt;
   *  Random number seed.
   *  (default 1)</pre>
   * 
   * <pre> -I &lt;num&gt;
   *  Number of iterations.
   *  (default 10)</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.DecisionStump)</pre>
   * 
   * <pre> 
   * Options specific to classifier weka.classifiers.trees.DecisionStump:
   * </pre>
   * 
   * <pre> -D
   *  If set, classifier is run in debug mode and
   *  may output additional info to the console</pre>
   * 
   <!-- options-end -->
   *
   * Options after -- are passed to the designated learner.<p>
   *
   * @param options the list of options as an array of strings
   * @throws Exception if an option is not supported
   */
  public void setOptions(String[] options) throws Exception {