makedensitybasedclusterer.java

数据挖掘中聚类的算法

/*
 *    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.
 */

/*
 *    MakeDensityBasedClusterer.java
 *    Copyright (C) 2002 University of Waikato, Hamilton, New Zealand
 *
 */

package weka.clusterers;

import weka.core.Capabilities;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.Option;
import weka.core.OptionHandler;
import weka.core.Utils;
import weka.core.WeightedInstancesHandler;
import weka.estimators.DiscreteEstimator;
import weka.filters.unsupervised.attribute.ReplaceMissingValues;

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

/**
 <!-- globalinfo-start -->
 * Class for wrapping a Clusterer to make it return a distribution and density. Fits normal distributions and discrete distributions within each cluster produced by the wrapped clusterer. Supports the NumberOfClustersRequestable interface only if the wrapped Clusterer does.
 * <p/>
 <!-- globalinfo-end -->
 *
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -M &lt;num&gt;
 *  minimum allowable standard deviation for normal density computation 
 *  (default 1e-6)</pre>
 * 
 * <pre> -W &lt;clusterer name&gt;
 *  Clusterer to wrap.
 *  (default weka.clusterers.SimpleKMeans)</pre>
 * 
 * <pre> 
 * Options specific to clusterer weka.clusterers.SimpleKMeans:
 * </pre>
 * 
 * <pre> -N &lt;num&gt;
 *  number of clusters. (default = 2).</pre>
 * 
 * <pre> -S &lt;num&gt;
 *  random number seed.
 *  (default 10)</pre>
 * 
 <!-- options-end -->
 * 
 * Options after "--" are passed on to the base clusterer.
 *
 * @author Richard Kirkby ([e-mail])
 * @author Mark Hall ([e-mail])
 * @author Eibe Frank ([e-mail])
 * @version $Revision: 1.13 $
 */
public class MakeDensityBasedClusterer 
  extends DensityBasedClusterer
  implements NumberOfClustersRequestable, 
	     OptionHandler, 
	     WeightedInstancesHandler {

  /** for serialization */
  static final long serialVersionUID = -5643302427972186631L;
  
  /** holds training instances header information */
  private Instances m_theInstances;
  /** prior probabilities for the fitted clusters */
  private double [] m_priors;
  /** normal distributions fitted to each numeric attribute in each cluster */
  private double [][][] m_modelNormal;
  /** discrete distributions fitted to each discrete attribute in each cluster */
  private DiscreteEstimator [][] m_model;
  /** default minimum standard deviation */
  private double m_minStdDev = 1e-6;
  /** The clusterer being wrapped */
  private Clusterer m_wrappedClusterer = new weka.clusterers.SimpleKMeans();
  /** globally replace missing values */
  private ReplaceMissingValues m_replaceMissing;

  /**
   * Default constructor.
   * 
   */  
  public MakeDensityBasedClusterer() {
    super();
  }
   
  /**
   * Contructs a MakeDensityBasedClusterer wrapping a given Clusterer.
   * 
   * @param toWrap the clusterer to wrap around
   */    
  public MakeDensityBasedClusterer(Clusterer toWrap) {

    setClusterer(toWrap);
  }
  
  /**
   * Returns a string describing classifier
   * @return a description suitable for
   * displaying in the explorer/experimenter gui
   */
  public String globalInfo() {
    return 
        "Class for wrapping a Clusterer to make it return a distribution "
      + "and density. Fits normal distributions and discrete distributions "
      + "within each cluster produced by the wrapped clusterer. Supports the "
      + "NumberOfClustersRequestable interface only if the wrapped Clusterer "
      + "does.";
  }

  /**
   * String describing default clusterer.
   * 
   * @return 		the default clusterer classname
   */
  protected String defaultClustererString() {
    return SimpleKMeans.class.getName();
  }

  /**
   * Set the number of clusters to generate.
   *
   * @param n the number of clusters to generate
   * @throws Exception if the wrapped clusterer has not been set, or if
   * the wrapped clusterer does not implement this facility.
   */
  public void setNumClusters(int n) throws Exception {
    if (m_wrappedClusterer == null) {
      throw new Exception("Can't set the number of clusters to generate - "
			  +"no clusterer has been set yet.");
    }
    if (!(m_wrappedClusterer instanceof NumberOfClustersRequestable)) {
      throw new Exception("Can't set the number of clusters to generate - "
			  +"wrapped clusterer does not support this facility.");
    }

    ((NumberOfClustersRequestable)m_wrappedClusterer).setNumClusters(n);
  }

  /**
   * Returns default capabilities of the clusterer (i.e., of the wrapper
   * clusterer).
   *
   * @return      the capabilities of this clusterer
   */
  public Capabilities getCapabilities() {
    if (m_wrappedClusterer != null)
      return m_wrappedClusterer.getCapabilities();
    else
      return super.getCapabilities();
  }
  
  /**
   * Builds a clusterer for a set of instances.
   *
   * @param data the instances to train the clusterer with
   * @throws Exception if the clusterer hasn't been set or something goes wrong
   */  
  public void buildClusterer(Instances data) throws Exception {
    // can clusterer handle the data?
    getCapabilities().testWithFail(data);

    m_replaceMissing = new ReplaceMissingValues();
    m_replaceMissing.setInputFormat(data);
    data = weka.filters.Filter.useFilter(data, m_replaceMissing);

    m_theInstances = new Instances(data, 0);
    if (m_wrappedClusterer == null) {
      throw new Exception("No clusterer has been set");
    }
    m_wrappedClusterer.buildClusterer(data);
    m_model = 
       new DiscreteEstimator[m_wrappedClusterer.numberOfClusters()][data.numAttributes()];
    m_modelNormal = 
      new double[m_wrappedClusterer.numberOfClusters()][data.numAttributes()][2];
    double[][] weights =  new double[m_wrappedClusterer.numberOfClusters()][data.numAttributes()];
    m_priors = new double[m_wrappedClusterer.numberOfClusters()]; 
     for (int i = 0; i < m_wrappedClusterer.numberOfClusters(); i++) {
       for (int j = 0; j < data.numAttributes(); j++) {
	 if (data.attribute(j).isNominal()) {
	   m_model[i][j] = new DiscreteEstimator(data.attribute(j).numValues(),
						 true);
	 }
       }
     }
     
     Instance inst = null;

     // Compute mean, etc.
     int[] clusterIndex = new int[data.numInstances()];
     for (int i = 0; i < data.numInstances(); i++) {
       inst = data.instance(i);
       int cluster = m_wrappedClusterer.clusterInstance(inst);
       m_priors[cluster] += inst.weight();
       for (int j = 0; j < data.numAttributes(); j++) {
	 if (!inst.isMissing(j)) {
	   if (data.attribute(j).isNominal()) {
	     m_model[cluster][j].addValue(inst.value(j),inst.weight());
	   } else {
	     m_modelNormal[cluster][j][0] += inst.weight() * inst.value(j);
	     weights[cluster][j] += inst.weight();
	   }
	 }
       }
       clusterIndex[i] = cluster;
     }

     for (int j = 0; j < data.numAttributes(); j++) {
       if (data.attribute(j).isNumeric()) {
	 for (int i = 0; i < m_wrappedClusterer.numberOfClusters(); i++) {	   
	   if (weights[i][j] > 0) {
	     m_modelNormal[i][j][0] /= weights[i][j];
	   }
	 }
       }
     }

     // Compute standard deviations
     for (int i = 0; i < data.numInstances(); i++) {
       inst = data.instance(i);
       for (int j = 0; j < data.numAttributes(); j++) {
	 if (!inst.isMissing(j)) {
	   if (data.attribute(j).isNumeric()) {
	     double diff = m_modelNormal[clusterIndex[i]][j][0] - inst.value(j);
	     m_modelNormal[clusterIndex[i]][j][1] += inst.weight() * diff * diff;
	   }
	 }
       }
     }

     for (int j = 0; j < data.numAttributes(); j++) {
       if (data.attribute(j).isNumeric()) {
	 for (int i = 0; i < m_wrappedClusterer.numberOfClusters(); i++) {	   
	   if (weights[i][j] > 0) {
	     m_modelNormal[i][j][1] = 
	       Math.sqrt(m_modelNormal[i][j][1] / weights[i][j]);
	   } else if (weights[i][j] <= 0) {
	     m_modelNormal[i][j][1] = Double.MAX_VALUE;
	   }
	   if (m_modelNormal[i][j][1] <= m_minStdDev) {
	     m_modelNormal[i][j][1] = data.attributeStats(j).numericStats.stdDev;
	     if (m_modelNormal[i][j][1] <= m_minStdDev) {
	       m_modelNormal[i][j][1] = m_minStdDev;
	     }
	   }
	 }
       }
     }
     
     Utils.normalize(m_priors);
  }

  /**
   * Returns the cluster priors.
   * 
   * @return the cluster priors
   */
  public double[] clusterPriors() {

    double[] n = new double[m_priors.length];
  
    System.arraycopy(m_priors, 0, n, 0, n.length);