citationknn.java

用于快速分类

/*
 *    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.
 */

/*
 * CitationKNN.java
 * Copyright (C) 2005 University of Waikato, Hamilton, New Zealand
 */

package weka.classifiers.mi;

import weka.classifiers.Classifier;
import weka.core.Capabilities;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.MultiInstanceCapabilitiesHandler;
import weka.core.Option;
import weka.core.OptionHandler;
import weka.core.TechnicalInformation;
import weka.core.TechnicalInformationHandler;
import weka.core.Utils;
import weka.core.Capabilities.Capability;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;

import java.io.Serializable;
import java.util.Enumeration;
import java.util.Vector;
/**
 <!-- globalinfo-start -->
 * Modified version of the Citation kNN multi instance classifier.<br/>
 * <br/>
 * For more information see:<br/>
 * <br/>
 * Jun Wang, Zucker, Jean-Daniel: Solving Multiple-Instance Problem: A Lazy Learning Approach. In: 17th International Conference on Machine Learning, 1119-1125, 2000.
 * <p/>
 <!-- globalinfo-end -->
 * 
 <!-- technical-bibtex-start -->
 * BibTeX:
 * <pre>
 * &#64;inproceedings{Wang2000,
 *    author = {Jun Wang and Zucker and Jean-Daniel},
 *    booktitle = {17th International Conference on Machine Learning},
 *    editor = {Pat Langley},
 *    pages = {1119-1125},
 *    title = {Solving Multiple-Instance Problem: A Lazy Learning Approach},
 *    year = {2000}
 * }
 * </pre>
 * <p/>
 <!-- technical-bibtex-end -->
 * 
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -R &lt;number of references&gt;
 *  Number of Nearest References (default 1)</pre>
 * 
 * <pre> -C &lt;number of citers&gt;
 *  Number of Nearest Citers (default 1)</pre>
 * 
 * <pre> -H &lt;rank&gt;
 *  Rank of the Hausdorff Distance (default 1)</pre>
 * 
 <!-- options-end -->
 *
 * @author Miguel Garcia Torres ([e-mail])
 * @version $Revision: 1.5 $ 
 */
public class CitationKNN 
  extends Classifier 
  implements OptionHandler, MultiInstanceCapabilitiesHandler,
             TechnicalInformationHandler {

  /** for serialization */
  static final long serialVersionUID = -8435377743874094852L;
  
  /** The index of the class attribute */
  protected int m_ClassIndex;

  /** The number of the class labels */
  protected int m_NumClasses;

  /** */
  protected int m_IdIndex;    

  /** Debugging output */
  protected boolean m_Debug;

  /** Class labels for each bag */
  protected int[] m_Classes;

  /** attribute name structure of the relational attribute*/
  protected Instances m_Attributes;

  /** Number of references */
  protected int m_NumReferences = 1;

  /** Number of citers*/
  protected int m_NumCiters = 1;

  /** Training bags*/
  protected Instances m_TrainBags;

  /** Different debugging output */
  protected boolean m_CNNDebug = false;

  protected boolean m_CitersDebug = false;

  protected boolean m_ReferencesDebug = false;

  protected boolean m_HDistanceDebug = false;

  protected boolean m_NeighborListDebug = false;

  /** C nearest neighbors considering all the bags*/
  protected NeighborList[] m_CNN;

  /** C nearest citers */
  protected int[] m_Citers;

  /** R nearest references */
  protected int[] m_References;

  /** Rank associated to the Hausdorff distance*/
  protected int m_HDRank = 1;

  /** Normalization of the euclidean distance */
  private double[] m_Diffs;

  private double[] m_Min;

  private double m_MinNorm = 0.95;

  private double[] m_Max;

  private double m_MaxNorm = 1.05;

  /**
   * Returns a string describing this filter
   *
   * @return a description of the filter suitable for
   * displaying in the explorer/experimenter gui
   */
  public String globalInfo() {
    return 
        "Modified version of the Citation kNN multi instance classifier.\n\n"
      + "For more information see:\n\n"
      + getTechnicalInformation().toString();
  }

  /**
   * 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.INPROCEEDINGS);
    result.setValue(Field.AUTHOR, "Jun Wang and Zucker and Jean-Daniel");
    result.setValue(Field.TITLE, "Solving Multiple-Instance Problem: A Lazy Learning Approach");
    result.setValue(Field.BOOKTITLE, "17th International Conference on Machine Learning");
    result.setValue(Field.EDITOR, "Pat Langley");
    result.setValue(Field.YEAR, "2000");
    result.setValue(Field.PAGES, "1119-1125");
    
    return result;
  }

  /** 
   * Calculates the normalization of each attribute.
   */
  public void preprocessData(){
    int i,j, k;
    double min, max;
    Instances instances;
    Instance instance;
    // compute the min/max of each feature

    for (i=0;i<m_Attributes.numAttributes();i++) {
      min=Double.POSITIVE_INFINITY ;
      max=Double.NEGATIVE_INFINITY ;
      for(j = 0; j < m_TrainBags.numInstances(); j++){
        instances = m_TrainBags.instance(j).relationalValue(1);
        for (k=0;k<instances.numInstances();k++) {
          instance = instances.instance(k);
          if(instance.value(i) < min)
            min= instance.value(i);
          if(instance.value(i) > max)
            max= instance.value(i);
        }
      }
      m_Min[i] = min * m_MinNorm;
      m_Max[i] = max * m_MaxNorm;
      m_Diffs[i]= max * m_MaxNorm - min * m_MinNorm;
    }	    

  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String HDRankTipText() {
    return "The rank associated to the Hausdorff distance.";
  }

  /**
   * Sets the rank associated to the Hausdorff distance
   * @param hDRank the rank of the Hausdorff distance
   */
  public void setHDRank(int hDRank){
    m_HDRank = hDRank;
  }

  /**
   * Returns the rank associated to the Hausdorff distance
   * @return the rank number
   */
  public int getHDRank(){
    return m_HDRank;
  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String numReferencesTipText() {
    return 
        "The number of references considered to estimate the class "
      + "prediction of tests bags.";
  }

  /**
   * Sets the number of references considered to estimate
   * the class prediction of tests bags
   * @param numReferences the number of references
   */
  public void setNumReferences(int numReferences){
    m_NumReferences = numReferences;
  }

  /**
   * Returns the number of references considered to estimate
   * the class prediction of tests bags
   * @return the number of references
   */
  public int getNumReferences(){
    return m_NumReferences;
  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String numCitersTipText() {
    return 
        "The number of citers considered to estimate the class "
      + "prediction of test bags.";
  }

  /**
   * Sets the number of citers considered to estimate
   * the class prediction of tests bags
   * @param numCiters the number of citers
   */
  public void setNumCiters(int numCiters){
    m_NumCiters = numCiters;
  }

  /**
   * Returns the number of citers considered to estimate
   * the class prediction of tests bags
   * @return the number of citers
   */
  public int getNumCiters(){
    return m_NumCiters;
  }

  /**
   * Returns default capabilities of the classifier.
   *
   * @return      the capabilities of this classifier
   */
  public Capabilities getCapabilities() {
    Capabilities result = super.getCapabilities();

    // attributes
    result.enable(Capability.NOMINAL_ATTRIBUTES);
    result.enable(Capability.NUMERIC_ATTRIBUTES);
    result.enable(Capability.DATE_ATTRIBUTES);
    result.enable(Capability.RELATIONAL_ATTRIBUTES);
    result.enable(Capability.MISSING_VALUES);

    // class
    result.enable(Capability.NOMINAL_CLASS);
    result.enable(Capability.MISSING_CLASS_VALUES);
    
    // other
    result.enable(Capability.ONLY_MULTIINSTANCE);
    
    return result;
  }

  /**
   * Returns the capabilities of this multi-instance classifier for the
   * relational data.
   *
   * @return            the capabilities of this object
   * @see               Capabilities
   */
  public Capabilities getMultiInstanceCapabilities() {
    Capabilities result = super.getCapabilities();
    
    // attributes
    result.enable(Capability.NOMINAL_ATTRIBUTES);
    result.enable(Capability.NUMERIC_ATTRIBUTES);
    result.enable(Capability.DATE_ATTRIBUTES);
    result.enable(Capability.MISSING_VALUES);

    // class
    result.disableAllClasses();
    result.enable(Capability.NO_CLASS);
    
    return result;
  }

  /**
   * Builds the classifier
   *
   * @param train the training data to be used for generating the
   * boosted classifier.
   * @throws Exception if the classifier could not be built successfully
   */
  public void buildClassifier(Instances train) throws Exception {
    // can classifier handle the data?
    getCapabilities().testWithFail(train);

    // remove instances with missing class