optics.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.clusterers.forOPTICSAndDBScan.OPTICS_GUI.OPTICS_Visualizer;
import weka.clusterers.forOPTICSAndDBScan.OPTICS_GUI.SERObject;
import weka.clusterers.forOPTICSAndDBScan.Utils.EpsilonRange_ListElement;
import weka.clusterers.forOPTICSAndDBScan.Utils.UpdateQueue;
import weka.clusterers.forOPTICSAndDBScan.Utils.UpdateQueueElement;
import weka.core.Capabilities;
import weka.core.FastVector;
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.io.BufferedWriter;
import java.io.FileWriter;
import java.lang.reflect.Constructor;
import java.lang.reflect.InvocationTargetException;
import java.text.DecimalFormat;
import java.util.Calendar;
import java.util.Enumeration;
import java.util.GregorianCalendar;
import java.util.Iterator;
import java.util.List;
import java.util.Vector;

/**
 <!-- globalinfo-start -->
 * Mihael Ankerst, Markus M. Breunig, Hans-Peter Kriegel, Joerg Sander: OPTICS: Ordering Points To Identify the Clustering Structure. In: ACM SIGMOD International Conference on Management of Data, 49-60, 1999.
 * <p/>
 <!-- globalinfo-end -->
 *
 <!-- technical-bibtex-start -->
 * BibTeX:
 * <pre>
 * &#64;inproceedings{Ankerst1999,
 *    author = {Mihael Ankerst and Markus M. Breunig and Hans-Peter Kriegel and Joerg Sander},
 *    booktitle = {ACM SIGMOD International Conference on Management of Data},
 *    pages = {49-60},
 *    publisher = {ACM Press},
 *    title = {OPTICS: Ordering Points To Identify the Clustering Structure},
 *    year = {1999}
 * }
 * </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 OPTICS (default = weka.clusterers.forOPTICSAndDBScan.Databases.SequentialDatabase)</pre>
 * 
 * <pre> -D &lt;String&gt;
 *  distance-type (default = weka.clusterers.forOPTICSAndDBScan.DataObjects.EuclidianDataObject)</pre>
 * 
 * <pre> -F
 *  write results to OPTICS_#TimeStamp#.TXT - File</pre>
 * 
 <!-- options-end -->
 *
 * @author Matthias Schubert ([e-mail])
 * @author Zhanna Melnikova-Albrecht ([e-mail])
 * @author Rainer Holzmann ([e-mail])
 * @version $Revision: 1.6 $
 */
public class OPTICS 
    extends Clusterer 
    implements OptionHandler, TechnicalInformationHandler {

    /** for serialization */
    static final long serialVersionUID = 274552680222105221L;
  
    /**
     * 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 OPTICS
     */
    private Database database;

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

    /**
     * Flag that indicates if the results are written to a file or not
     */
    private boolean writeOPTICSresults = false;

    /**
     * Holds the ClusterOrder (dataObjects with their r_dist and c_dist) for the GUI
     */
    private FastVector resultVector;

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

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

    /**
     * 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;
    }

    /**
     * Generate Clustering via OPTICS
     * @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);

        resultVector = new FastVector();
        long time_1 = System.currentTimeMillis();

        numberOfGeneratedClusters = 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();

        UpdateQueue seeds = new UpdateQueue();

        /** OPTICS-Begin */
        Iterator iterator = database.dataObjectIterator();
        while (iterator.hasNext()) {
            DataObject dataObject = (DataObject) iterator.next();
            if (!dataObject.isProcessed()) {
                expandClusterOrder(dataObject, seeds);
            }
        }

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

        if (writeOPTICSresults) {
            String fileName = "";
            GregorianCalendar gregorianCalendar = new GregorianCalendar();
            String timeStamp = gregorianCalendar.get(Calendar.DAY_OF_MONTH) + "-" +
                    (gregorianCalendar.get(Calendar.MONTH) + 1) +
                    "-" + gregorianCalendar.get(Calendar.YEAR) +
                    "--" + gregorianCalendar.get(Calendar.HOUR_OF_DAY) +
                    "-" + gregorianCalendar.get(Calendar.MINUTE) +
                    "-" + gregorianCalendar.get(Calendar.SECOND);
            fileName = "OPTICS_" + timeStamp + ".TXT";

            FileWriter fileWriter = new FileWriter(fileName);
            BufferedWriter bufferedOPTICSWriter = new BufferedWriter(fileWriter);
            for (int i = 0; i < resultVector.size(); i++) {
                bufferedOPTICSWriter.write(format_dataObject((DataObject) resultVector.elementAt(i)));
            }
            bufferedOPTICSWriter.flush();
            bufferedOPTICSWriter.close();
        }

        new OPTICS_Visualizer(getSERObject(), "OPTICS Visualizer - Main Window");
    }

    /**
     * Expands the ClusterOrder for this dataObject
     * @param dataObject Start-DataObject
     * @param seeds SeedList that stores dataObjects with reachability-distances
     */
    private void expandClusterOrder(DataObject dataObject, UpdateQueue seeds) {
        List list = database.coreDistance(getMinPoints(), getEpsilon(), dataObject);
        List epsilonRange_List = (List) list.get(1);
        dataObject.setReachabilityDistance(DataObject.UNDEFINED);
        dataObject.setCoreDistance(((Double) list.get(2)).doubleValue());
        dataObject.setProcessed(true);

        resultVector.addElement(dataObject);

        if (dataObject.getCoreDistance() != DataObject.UNDEFINED) {
            update(seeds, epsilonRange_List, dataObject);
            while (seeds.hasNext()) {
                UpdateQueueElement updateQueueElement = seeds.next();
                DataObject currentDataObject = (DataObject) updateQueueElement.getObject();
                currentDataObject.setReachabilityDistance(updateQueueElement.getPriority());
                List list_1 = database.coreDistance(getMinPoints(), getEpsilon(), currentDataObject);
                List epsilonRange_List_1 = (List) list_1.get(1);
                currentDataObject.setCoreDistance(((Double) list_1.get(2)).doubleValue());
                currentDataObject.setProcessed(true);

                resultVector.addElement(currentDataObject);

                if (currentDataObject.getCoreDistance() != DataObject.UNDEFINED) {
                    update(seeds, epsilonRange_List_1, currentDataObject);
                }
            }
        }
    }

    /**
     * Wraps the dataObject into a String, that contains the dataObject's key, the dataObject itself,
     * the coreDistance and its reachabilityDistance in a formatted manner.
     * @param dataObject The dataObject that is wrapped into a formatted string.
     * @return String Formatted string
     */
    private String format_dataObject(DataObject dataObject) {
        StringBuffer stringBuffer = new StringBuffer();

        stringBuffer.append("(" + Utils.doubleToString(Double.parseDouble(dataObject.getKey()),
                (Integer.toString(database.size()).length()), 0) + ".) "
                + Utils.padRight(dataObject.toString(), 40) + "  -->  c_dist: " +

                ((dataObject.getCoreDistance() == DataObject.UNDEFINED) ?
                Utils.padRight("UNDEFINED", 12) :
                Utils.padRight(Utils.doubleToString(dataObject.getCoreDistance(), 2, 3), 12)) +

                " r_dist: " +
                ((dataObject.getReachabilityDistance() == DataObject.UNDEFINED) ?
                Utils.padRight("UNDEFINED", 12) :
                Utils.doubleToString(dataObject.getReachabilityDistance(), 2, 3)) + "\n");

        return stringBuffer.toString();
    }

    /**
     * Updates reachability-distances in the Seeds-List
     * @param seeds UpdateQueue that holds DataObjects with their corresponding reachability-distances
     * @param epsilonRange_list List of DataObjects that were found in epsilon-range of centralObject
     * @param centralObject
     */
    private void update(UpdateQueue seeds, List epsilonRange_list, DataObject centralObject) {
        double coreDistance = centralObject.getCoreDistance();
        double new_r_dist = DataObject.UNDEFINED;

        for (int i = 0; i < epsilonRange_list.size(); i++) {
            EpsilonRange_ListElement listElement = (EpsilonRange_ListElement) epsilonRange_list.get(i);
            DataObject neighbourhood_object = listElement.getDataObject();
            if (!neighbourhood_object.isProcessed()) {
                new_r_dist = Math.max(coreDistance, listElement.getDistance());
                seeds.add(new_r_dist, neighbourhood_object, neighbourhood_object.getKey());
            }
        }
    }

    /**
     * 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 {
        throw new Exception();
    }

    /**
     * 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 OPTICS (default = weka.clusterers.forOPTICSAndDBScan.Databases.SequentialDatabase)",
                        "I",
                        1,
                        "-I <String>"));
        vector.addElement(