dbscan.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.
 */

/*
 *    Copyright (C) 2004
 *    & Matthias Schubert ([e-mail])
 *    & Zhanna Melnikova-Albrecht ([e-mail])
 *    & Rainer Holzmann ([e-mail])
 */

package weka.clusterers;

import weka.clusterers.forOPTICSAndDBScan.DataObjects.DataObject;
import weka.clusterers.forOPTICSAndDBScan.Databases.Database;
import weka.core.Capabilities;
import weka.core.Instance;
import weka.core.Instances;
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 weka.filters.Filter;
import weka.filters.unsupervised.attribute.ReplaceMissingValues;

import java.lang.reflect.Constructor;
import java.lang.reflect.InvocationTargetException;
import java.text.DecimalFormat;
import java.util.Enumeration;
import java.util.Iterator;
import java.util.List;
import java.util.Vector;

/**
 <!-- globalinfo-start -->
 * Martin Ester, Hans-Peter Kriegel, Joerg Sander, Xiaowei Xu: A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise. In: Second International Conference on Knowledge Discovery and Data Mining, 226-231, 1996.
 * <p/>
 <!-- globalinfo-end -->
 *
 <!-- technical-bibtex-start -->
 * BibTeX:
 * <pre>
 * &#64;inproceedings{Ester1996,
 *    author = {Martin Ester and Hans-Peter Kriegel and Joerg Sander and Xiaowei Xu},
 *    booktitle = {Second International Conference on Knowledge Discovery and Data Mining},
 *    editor = {Evangelos Simoudis and Jiawei Han and Usama M. Fayyad},
 *    pages = {226-231},
 *    publisher = {AAAI Press},
 *    title = {A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise},
 *    year = {1996}
 * }
 * </pre>
 * <p/>
 <!-- technical-bibtex-end -->
 *
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -E &lt;double&gt;
 *  epsilon (default = 0.9)</pre>
 * 
 * <pre> -M &lt;int&gt;
 *  minPoints (default = 6)</pre>
 * 
 * <pre> -I &lt;String&gt;
 *  index (database) used for DBScan (default = weka.clusterers.forOPTICSAndDBScan.Databases.SequentialDatabase)</pre>
 * 
 * <pre> -D &lt;String&gt;
 *  distance-type (default = weka.clusterers.forOPTICSAndDBScan.DataObjects.EuclidianDataObject)</pre>
 * 
 <!-- options-end -->
 *
 * @author Matthias Schubert ([e-mail])
 * @author Zhanna Melnikova-Albrecht ([e-mail])
 * @author Rainer Holzmann ([e-mail])
 * @version $Revision: 1.7 $
 */
public class DBScan 
    extends Clusterer 
    implements OptionHandler, TechnicalInformationHandler {

    /** for serialization */
    static final long serialVersionUID = -1666498248451219728L;
  
    /**
     * Specifies the radius for a range-query
     */
    private double epsilon = 0.9;

    /**
     * Specifies the density (the range-query must contain at least minPoints DataObjects)
     */
    private int minPoints = 6;

    /**
     * Replace missing values in training instances
     */
    private ReplaceMissingValues replaceMissingValues_Filter;

    /**
     * Holds the number of clusters generated
     */
    private int numberOfGeneratedClusters;

    /**
     * Holds the distance-type that is used
     * (default = weka.clusterers.forOPTICSAndDBScan.DataObjects.EuclidianDataObject)
     */
    private String database_distanceType = "weka.clusterers.forOPTICSAndDBScan.DataObjects.EuclidianDataObject";

    /**
     * Holds the type of the used database
     * (default = weka.clusterers.forOPTICSAndDBScan.Databases.SequentialDatabase)
     */
    private String database_Type = "weka.clusterers.forOPTICSAndDBScan.Databases.SequentialDatabase";

    /**
     * The database that is used for DBScan
     */
    private Database database;

    /**
     * Holds the current clusterID
     */
    private int clusterID;

    /**
     * Counter for the processed instances
     */
    private int processed_InstanceID;

    /**
     * Holds the time-value (seconds) for the duration of the clustering-process
     */
    private double elapsedTime;

    /**
     * Returns default capabilities of the clusterer.
     *
     * @return      the capabilities of this clusterer
     */
    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.MISSING_VALUES);

      return result;
    }

    // *****************************************************************************************************************
    // constructors
    // *****************************************************************************************************************

    // *****************************************************************************************************************
    // methods
    // *****************************************************************************************************************

    /**
     * Generate Clustering via DBScan
     * @param instances The instances that need to be clustered
     * @throws java.lang.Exception If clustering was not successful
     */
    public void buildClusterer(Instances instances) throws Exception {
        // can clusterer handle the data?
        getCapabilities().testWithFail(instances);

        long time_1 = System.currentTimeMillis();

        processed_InstanceID = 0;
        numberOfGeneratedClusters = 0;
        clusterID = 0;

        replaceMissingValues_Filter = new ReplaceMissingValues();
        replaceMissingValues_Filter.setInputFormat(instances);
        Instances filteredInstances = Filter.useFilter(instances, replaceMissingValues_Filter);

        database = databaseForName(getDatabase_Type(), filteredInstances);
        for (int i = 0; i < database.getInstances().numInstances(); i++) {
            DataObject dataObject = dataObjectForName(getDatabase_distanceType(),
                    database.getInstances().instance(i),
                    Integer.toString(i),
                    database);
            database.insert(dataObject);
        }
        database.setMinMaxValues();

        Iterator iterator = database.dataObjectIterator();
        while (iterator.hasNext()) {
            DataObject dataObject = (DataObject) iterator.next();
            if (dataObject.getClusterLabel() == DataObject.UNCLASSIFIED) {
                if (expandCluster(dataObject)) {
                    clusterID++;
                    numberOfGeneratedClusters++;
                }
            }
        }

        long time_2 = System.currentTimeMillis();
        elapsedTime = (double) (time_2 - time_1) / 1000.0;
    }

    /**
     * Assigns this dataObject to a cluster or remains it as NOISE
     * @param dataObject The DataObject that needs to be assigned
     * @return true, if the DataObject could be assigned, else false
     */
    private boolean expandCluster(DataObject dataObject) {
        List seedList = database.epsilonRangeQuery(getEpsilon(), dataObject);
        /** dataObject is NO coreObject */
        if (seedList.size() < getMinPoints()) {
            dataObject.setClusterLabel(DataObject.NOISE);
            return false;
        }

        /** dataObject is coreObject */
        for (int i = 0; i < seedList.size(); i++) {
            DataObject seedListDataObject = (DataObject) seedList.get(i);
            /** label this seedListDataObject with the current clusterID, because it is in epsilon-range */
            seedListDataObject.setClusterLabel(clusterID);
            if (seedListDataObject.equals(dataObject)) {
                seedList.remove(i);
                i--;
            }
        }

        /** Iterate the seedList of the startDataObject */
        for (int j = 0; j < seedList.size(); j++) {
            DataObject seedListDataObject = (DataObject) seedList.get(j);
            List seedListDataObject_Neighbourhood = database.epsilonRangeQuery(getEpsilon(), seedListDataObject);

            /** seedListDataObject is coreObject */
            if (seedListDataObject_Neighbourhood.size() >= getMinPoints()) {
                for (int i = 0; i < seedListDataObject_Neighbourhood.size(); i++) {
                    DataObject p = (DataObject) seedListDataObject_Neighbourhood.get(i);
                    if (p.getClusterLabel() == DataObject.UNCLASSIFIED || p.getClusterLabel() == DataObject.NOISE) {
                        if (p.getClusterLabel() == DataObject.UNCLASSIFIED) {
                            seedList.add(p);
                        }
                        p.setClusterLabel(clusterID);
                    }
                }
            }
            seedList.remove(j);
            j--;
        }

        return true;
    }

    /**
     * Classifies a given instance.
     *
     * @param instance The instance to be assigned to a cluster
     * @return int The number of the assigned cluster as an integer
     * @throws java.lang.Exception If instance could not be clustered
     * successfully
     */
    public int clusterInstance(Instance instance) throws Exception {
        if (processed_InstanceID >= database.size()) processed_InstanceID = 0;
        int cnum = (database.getDataObject(Integer.toString(processed_InstanceID++))).getClusterLabel();
        if (cnum == DataObject.NOISE)
            throw new Exception();
        else
            return cnum;
    }

    /**
     * Returns the number of clusters.
     *
     * @return int The number of clusters generated for a training dataset.
     * @throws java.lang.Exception if number of clusters could not be returned
     * successfully
     */
    public int numberOfClusters() throws Exception {
        return numberOfGeneratedClusters;
    }

    /**
     * Returns an enumeration of all the available options..
     *
     * @return Enumeration An enumeration of all available options.
     */
    public Enumeration listOptions() {
        Vector vector = new Vector();

        vector.addElement(
                new Option("\tepsilon (default = 0.9)",
                        "E",
                        1,
                        "-E <double>"));
        vector.addElement(
                new Option("\tminPoints (default = 6)",
                        "M",
                        1,
                        "-M <int>"));
        vector.addElement(
                new Option("\tindex (database) used for DBScan (default = weka.clusterers.forOPTICSAndDBScan.Databases.SequentialDatabase)",
                        "I",