Package weka.gui.explorer

Source Code of weka.gui.explorer.ClustererPanel

/*
*    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.
*/

/*
*    ClustererPanel.java
*    Copyright (C) 1999 University of Waikato, Hamilton, New Zealand
*
*/

package weka.gui.explorer;

import weka.clusterers.ClusterEvaluation;
import weka.clusterers.Clusterer;
import weka.core.Attribute;
import weka.core.Capabilities;
import weka.core.CapabilitiesHandler;
import weka.core.Drawable;
import weka.core.FastVector;
import weka.core.Instances;
import weka.core.OptionHandler;
import weka.core.SerializedObject;
import weka.core.Utils;
import weka.core.Version;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.Remove;
import weka.gui.ExtensionFileFilter;
import weka.gui.GenericObjectEditor;
import weka.gui.InstancesSummaryPanel;
import weka.gui.ListSelectorDialog;
import weka.gui.Logger;
import weka.gui.PropertyPanel;
import weka.gui.ResultHistoryPanel;
import weka.gui.SaveBuffer;
import weka.gui.SetInstancesPanel;
import weka.gui.SysErrLog;
import weka.gui.TaskLogger;
import weka.gui.explorer.Explorer.CapabilitiesFilterChangeEvent;
import weka.gui.explorer.Explorer.CapabilitiesFilterChangeListener;
import weka.gui.explorer.Explorer.ExplorerPanel;
import weka.gui.explorer.Explorer.LogHandler;
import weka.gui.treevisualizer.PlaceNode2;
import weka.gui.treevisualizer.TreeVisualizer;
import weka.gui.visualize.VisualizePanel;
import weka.gui.visualize.plugins.TreeVisualizePlugin;

import java.awt.BorderLayout;
import java.awt.Dimension;
import java.awt.Font;
import java.awt.GridBagConstraints;
import java.awt.GridBagLayout;
import java.awt.GridLayout;
import java.awt.Insets;
import java.awt.Point;
import java.awt.event.ActionEvent;
import java.awt.event.ActionListener;
import java.awt.event.InputEvent;
import java.awt.event.MouseAdapter;
import java.awt.event.MouseEvent;
import java.beans.PropertyChangeEvent;
import java.beans.PropertyChangeListener;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.io.InputStream;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.io.OutputStream;
import java.text.SimpleDateFormat;
import java.util.Date;
import java.util.Random;
import java.util.Vector;
import java.util.zip.GZIPInputStream;
import java.util.zip.GZIPOutputStream;

import javax.swing.BorderFactory;
import javax.swing.ButtonGroup;
import javax.swing.DefaultComboBoxModel;
import javax.swing.DefaultListModel;
import javax.swing.JButton;
import javax.swing.JCheckBox;
import javax.swing.JComboBox;
import javax.swing.JFileChooser;
import javax.swing.JFrame;
import javax.swing.JLabel;
import javax.swing.JList;
import javax.swing.JMenu;
import javax.swing.JMenuItem;
import javax.swing.JOptionPane;
import javax.swing.JPanel;
import javax.swing.JPopupMenu;
import javax.swing.JRadioButton;
import javax.swing.JScrollPane;
import javax.swing.JTextArea;
import javax.swing.JTextField;
import javax.swing.JViewport;
import javax.swing.SwingConstants;
import javax.swing.event.ChangeEvent;
import javax.swing.event.ChangeListener;
import javax.swing.filechooser.FileFilter;

import weka.gui.hierarchyvisualizer.HierarchyVisualizer;

/**
* This panel allows the user to select and configure a clusterer, and evaluate
* the clusterer using a number of testing modes (test on the training data,
* train/test on a percentage split, test on a
* separate split). The results of clustering runs are stored in a result
* history so that previous results are accessible.
*
* @author Mark Hall (mhall@cs.waikato.ac.nz)
* @author Richard Kirkby (rkirkby@cs.waikato.ac.nz)
* @version $Revision: 6889 $
*/
public class ClustererPanel
  extends JPanel
  implements CapabilitiesFilterChangeListener, ExplorerPanel, LogHandler {

  /** for serialization */
  static final long serialVersionUID = -2474932792950820990L;

  /** the parent frame */
  protected Explorer m_Explorer = null;
 
  /** The filename extension that should be used for model files */
  public static String MODEL_FILE_EXTENSION = ".model";

  /** Lets the user configure the clusterer */
  protected GenericObjectEditor m_ClustererEditor =
    new GenericObjectEditor();

  /** The panel showing the current clusterer selection */
  protected PropertyPanel m_CLPanel = new PropertyPanel(m_ClustererEditor);
 
  /** The output area for classification results */
  protected JTextArea m_OutText = new JTextArea(20, 40);

  /** The destination for log/status messages */
  protected Logger m_Log = new SysErrLog();

  /** The buffer saving object for saving output */
  SaveBuffer m_SaveOut = new SaveBuffer(m_Log, this);

  /** A panel controlling results viewing */
  protected ResultHistoryPanel m_History = new ResultHistoryPanel(m_OutText);

  /** Click to set test mode to generate a % split */
  protected JRadioButton m_PercentBut = new JRadioButton("Percentage split");

  /** Click to set test mode to test on training data */
  protected JRadioButton m_TrainBut = new JRadioButton("Use training set");

  /** Click to set test mode to a user-specified test set */
  protected JRadioButton m_TestSplitBut =
    new JRadioButton("Supplied test set");

  /** Click to set test mode to classes to clusters based evaluation */
  protected JRadioButton m_ClassesToClustersBut =
    new JRadioButton("Classes to clusters evaluation");

  /** Lets the user select the class column for classes to clusters based
      evaluation */
  protected JComboBox m_ClassCombo = new JComboBox();

  /** Label by where the % split is entered */
  protected JLabel m_PercentLab = new JLabel("%", SwingConstants.RIGHT);

  /** The field where the % split is entered */
  protected JTextField m_PercentText = new JTextField("66");

  /** The button used to open a separate test dataset */
  protected JButton m_SetTestBut = new JButton("Set...");

  /** The frame used to show the test set selection panel */
  protected JFrame m_SetTestFrame;

  /** The button used to popup a list for choosing attributes to ignore while
      clustering */
  protected JButton m_ignoreBut = new JButton("Ignore attributes");

  protected DefaultListModel m_ignoreKeyModel = new DefaultListModel();
  protected JList m_ignoreKeyList = new JList(m_ignoreKeyModel);

  //  protected Remove m_ignoreFilter = null;
 
  /**
   * Alters the enabled/disabled status of elements associated with each
   * radio button
   */
  ActionListener m_RadioListener = new ActionListener() {
    public void actionPerformed(ActionEvent e) {
      updateRadioLinks();
    }
  };

  /** Click to start running the clusterer */
  protected JButton m_StartBut = new JButton("Start");

  /** Stop the class combo from taking up to much space */
  private Dimension COMBO_SIZE = new Dimension(250, m_StartBut
                 .getPreferredSize().height);

  /** Click to stop a running clusterer */
  protected JButton m_StopBut = new JButton("Stop");

  /** The main set of instances we're playing with */
  protected Instances m_Instances;

  /** The user-supplied test set (if any) */
  protected Instances m_TestInstances;

  /** The current visualization object */
  protected VisualizePanel m_CurrentVis = null;

  /** Check to save the predictions in the results list for visualizing
      later on */
  protected JCheckBox m_StorePredictionsBut =
    new JCheckBox("Store clusters for visualization");
 
  /** A thread that clustering runs in */
  protected Thread m_RunThread;
 
  /** The instances summary panel displayed by m_SetTestFrame */
  protected InstancesSummaryPanel m_Summary;

  /** Filter to ensure only model files are selected */ 
  protected FileFilter m_ModelFilter =
    new ExtensionFileFilter(MODEL_FILE_EXTENSION, "Model object files");

  /** The file chooser for selecting model files */
  protected JFileChooser m_FileChooser
    = new JFileChooser(new File(System.getProperty("user.dir")));

  /* Register the property editors we need */
  static {
     GenericObjectEditor.registerEditors();
  }
 
  /**
   * Creates the clusterer panel
   */
  public ClustererPanel() {

    // Connect / configure the components
    m_OutText.setEditable(false);
    m_OutText.setFont(new Font("Monospaced", Font.PLAIN, 12));
    m_OutText.setBorder(BorderFactory.createEmptyBorder(5, 5, 5, 5));
    m_OutText.addMouseListener(new MouseAdapter() {
      public void mouseClicked(MouseEvent e) {
  if ((e.getModifiers() & InputEvent.BUTTON1_MASK)
      != InputEvent.BUTTON1_MASK) {
    m_OutText.selectAll();
  }
      }
    });
    m_History.setBorder(BorderFactory.createTitledBorder("Result list (right-click for options)"));
    m_ClustererEditor.setClassType(Clusterer.class);
    m_ClustererEditor.setValue(ExplorerDefaults.getClusterer());
    m_ClustererEditor.addPropertyChangeListener(new PropertyChangeListener() {
      public void propertyChange(PropertyChangeEvent e) {
        m_StartBut.setEnabled(true);
        Capabilities currentFilter = m_ClustererEditor.getCapabilitiesFilter();
        Clusterer clusterer = (Clusterer) m_ClustererEditor.getValue();
        Capabilities currentSchemeCapabilities =  null;
        if (clusterer != null && currentFilter != null &&
            (clusterer instanceof CapabilitiesHandler)) {
          currentSchemeCapabilities = ((CapabilitiesHandler)clusterer).getCapabilities();
         
          if (!currentSchemeCapabilities.supportsMaybe(currentFilter) &&
              !currentSchemeCapabilities.supports(currentFilter)) {
            m_StartBut.setEnabled(false);
          }
        }
  repaint();
      }
    });

    m_TrainBut.setToolTipText("Cluster the same set that the clusterer"
            + " is trained on");
    m_PercentBut.setToolTipText("Train on a percentage of the data and"
        + " cluster the remainder");
    m_TestSplitBut.setToolTipText("Cluster a user-specified dataset");
    m_ClassesToClustersBut.setToolTipText("Evaluate clusters with respect to a"
            +" class");
    m_ClassCombo.setToolTipText("Select the class attribute for class based"
        +" evaluation");
    m_StartBut.setToolTipText("Starts the clustering");
    m_StopBut.setToolTipText("Stops a running clusterer");
    m_StorePredictionsBut.
      setToolTipText("Store predictions in the result list for later "
         +"visualization");
    m_ignoreBut.setToolTipText("Ignore attributes during clustering");

    m_FileChooser.setFileFilter(m_ModelFilter);
    m_FileChooser.setFileSelectionMode(JFileChooser.FILES_ONLY);

    m_ClassCombo.setPreferredSize(COMBO_SIZE);
    m_ClassCombo.setMaximumSize(COMBO_SIZE);
    m_ClassCombo.setMinimumSize(COMBO_SIZE);
    m_ClassCombo.setEnabled(false);

    m_PercentBut.setSelected(ExplorerDefaults.getClustererTestMode() == 2);
    m_TrainBut.setSelected(ExplorerDefaults.getClustererTestMode() == 3);
    m_TestSplitBut.setSelected(ExplorerDefaults.getClustererTestMode() == 4);
    m_ClassesToClustersBut.setSelected(ExplorerDefaults.getClustererTestMode() == 5);
    m_StorePredictionsBut.setSelected(ExplorerDefaults.getClustererStoreClustersForVis());
    updateRadioLinks();
    ButtonGroup bg = new ButtonGroup();
    bg.add(m_TrainBut);
    bg.add(m_PercentBut);
    bg.add(m_TestSplitBut);
    bg.add(m_ClassesToClustersBut);
    m_TrainBut.addActionListener(m_RadioListener);
    m_PercentBut.addActionListener(m_RadioListener);
    m_TestSplitBut.addActionListener(m_RadioListener);
    m_ClassesToClustersBut.addActionListener(m_RadioListener);
    m_SetTestBut.addActionListener(new ActionListener() {
      public void actionPerformed(ActionEvent e) {
  setTestSet();
      }
    });

    m_StartBut.setEnabled(false);
    m_StopBut.setEnabled(false);
    m_ignoreBut.setEnabled(false);
    m_StartBut.addActionListener(new ActionListener() {
      public void actionPerformed(ActionEvent e) {
  startClusterer();
      }
    });
    m_StopBut.addActionListener(new ActionListener() {
      public void actionPerformed(ActionEvent e) {
  stopClusterer();
      }
    });

    m_ignoreBut.addActionListener(new ActionListener() {
  public void actionPerformed(ActionEvent e) {
    setIgnoreColumns();
  }
      });
  
    m_History.setHandleRightClicks(false);
    // see if we can popup a menu for the selected result
    m_History.getList().addMouseListener(new MouseAdapter() {
  public void mouseClicked(MouseEvent e) {
    if (((e.getModifiers() & InputEvent.BUTTON1_MASK)
         != InputEvent.BUTTON1_MASK) || e.isAltDown()) {
      int index = m_History.getList().locationToIndex(e.getPoint());
      if (index != -1) {
        String name = m_History.getNameAtIndex(index);
        visualizeClusterer(name, e.getX(), e.getY());
      } else {
        visualizeClusterer(null, e.getX(), e.getY());
      }
    }
  }
      });
   
    m_ClassCombo.addActionListener(new ActionListener() {
      public void actionPerformed(ActionEvent e) {
  updateCapabilitiesFilter(m_ClustererEditor.getCapabilitiesFilter());
      }
    });

    // Layout the GUI
    JPanel p1 = new JPanel();
    p1.setBorder(BorderFactory.createCompoundBorder(
     BorderFactory.createTitledBorder("Clusterer"),
     BorderFactory.createEmptyBorder(0, 5, 5, 5)
     ));
    p1.setLayout(new BorderLayout());
    p1.add(m_CLPanel, BorderLayout.NORTH);

    JPanel p2 = new JPanel();
    GridBagLayout gbL = new GridBagLayout();
    p2.setLayout(gbL);
    p2.setBorder(BorderFactory.createCompoundBorder(
     BorderFactory.createTitledBorder("Cluster mode"),
     BorderFactory.createEmptyBorder(0, 5, 5, 5)
     ));
    GridBagConstraints gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.WEST;
    gbC.gridy = 0;     gbC.gridx = 0;
    gbL.setConstraints(m_TrainBut, gbC);
    p2.add(m_TrainBut);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.WEST;
    gbC.gridy = 1;     gbC.gridx = 0;
    gbL.setConstraints(m_TestSplitBut, gbC);
    p2.add(m_TestSplitBut);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.EAST;
    gbC.fill = GridBagConstraints.HORIZONTAL;
    gbC.gridy = 1;     gbC.gridx = 1;    gbC.gridwidth = 2;
    gbC.insets = new Insets(2, 10, 2, 0);
    gbL.setConstraints(m_SetTestBut, gbC);
    p2.add(m_SetTestBut);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.WEST;
    gbC.gridy = 2;     gbC.gridx = 0;
    gbL.setConstraints(m_PercentBut, gbC);
    p2.add(m_PercentBut);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.EAST;
    gbC.fill = GridBagConstraints.HORIZONTAL;
    gbC.gridy = 2;     gbC.gridx = 1;
    gbC.insets = new Insets(2, 10, 2, 10);
    gbL.setConstraints(m_PercentLab, gbC);
    p2.add(m_PercentLab);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.EAST;
    gbC.fill = GridBagConstraints.HORIZONTAL;
    gbC.gridy = 2;     gbC.gridx = 2;  gbC.weightx = 100;
    gbC.ipadx = 20;
    gbL.setConstraints(m_PercentText, gbC);
    p2.add(m_PercentText);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.WEST;
    gbC.gridy = 3;     gbC.gridx = 0;  gbC.gridwidth = 2;
    gbL.setConstraints(m_ClassesToClustersBut, gbC);
    p2.add(m_ClassesToClustersBut);

    m_ClassCombo.setBorder(BorderFactory.createEmptyBorder(0, 20, 0, 0));
    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.WEST;
    gbC.gridy = 4;     gbC.gridx = 0;  gbC.gridwidth = 2;
    gbL.setConstraints(m_ClassCombo, gbC);
    p2.add(m_ClassCombo);

    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.WEST;
    gbC.gridy = 5;     gbC.gridx = 0;  gbC.gridwidth = 2;
    gbL.setConstraints(m_StorePredictionsBut, gbC);
    p2.add(m_StorePredictionsBut);

    JPanel buttons = new JPanel();
    buttons.setLayout(new GridLayout(2, 1));
    JPanel ssButs = new JPanel();
    ssButs.setBorder(BorderFactory.createEmptyBorder(5, 5, 5, 5));
    ssButs.setLayout(new GridLayout(1, 2, 5, 5));
    ssButs.add(m_StartBut);
    ssButs.add(m_StopBut);

    JPanel ib = new JPanel();
    ib.setBorder(BorderFactory.createEmptyBorder(5, 5, 5, 5));
    ib.setLayout(new GridLayout(1, 1, 5, 5));
    ib.add(m_ignoreBut);
    buttons.add(ib);
    buttons.add(ssButs);
   
    JPanel p3 = new JPanel();
    p3.setBorder(BorderFactory.createTitledBorder("Clusterer output"));
    p3.setLayout(new BorderLayout());
    final JScrollPane js = new JScrollPane(m_OutText);
    p3.add(js, BorderLayout.CENTER);
    js.getViewport().addChangeListener(new ChangeListener() {
      private int lastHeight;
      public void stateChanged(ChangeEvent e) {
  JViewport vp = (JViewport)e.getSource();
  int h = vp.getViewSize().height;
  if (h != lastHeight) { // i.e. an addition not just a user scrolling
    lastHeight = h;
    int x = h - vp.getExtentSize().height;
    vp.setViewPosition(new Point(0, x));
  }
      }
    });   

    JPanel mondo = new JPanel();
    gbL = new GridBagLayout();
    mondo.setLayout(gbL);
    gbC = new GridBagConstraints();
    //    gbC.anchor = GridBagConstraints.WEST;
    gbC.fill = GridBagConstraints.HORIZONTAL;
    gbC.gridy = 0;     gbC.gridx = 0;
    gbL.setConstraints(p2, gbC);
    mondo.add(p2);
    gbC = new GridBagConstraints();
    gbC.anchor = GridBagConstraints.NORTH;
    gbC.fill = GridBagConstraints.HORIZONTAL;
    gbC.gridy = 1;     gbC.gridx = 0;
    gbL.setConstraints(buttons, gbC);
    mondo.add(buttons);
    gbC = new GridBagConstraints();
    //gbC.anchor = GridBagConstraints.NORTH;
    gbC.fill = GridBagConstraints.BOTH;
    gbC.gridy = 2;     gbC.gridx = 0; gbC.weightx = 0;
    gbL.setConstraints(m_History, gbC);
    mondo.add(m_History);
    gbC = new GridBagConstraints();
    gbC.fill = GridBagConstraints.BOTH;
    gbC.gridy = 0;     gbC.gridx = 1;
    gbC.gridheight = 3;
    gbC.weightx = 100; gbC.weighty = 100;
    gbL.setConstraints(p3, gbC);
    mondo.add(p3);

    setLayout(new BorderLayout());
    add(p1, BorderLayout.NORTH);
    add(mondo, BorderLayout.CENTER);
  }
 
  /**
   * Updates the enabled status of the input fields and labels.
   */
  protected void updateRadioLinks() {
   
    m_SetTestBut.setEnabled(m_TestSplitBut.isSelected());
    if ((m_SetTestFrame != null) && (!m_TestSplitBut.isSelected())) {
      m_SetTestFrame.setVisible(false);
    }
    m_PercentText.setEnabled(m_PercentBut.isSelected());
    m_PercentLab.setEnabled(m_PercentBut.isSelected());
    m_ClassCombo.setEnabled(m_ClassesToClustersBut.isSelected());
  }

  /**
   * Sets the Logger to receive informational messages
   *
   * @param newLog the Logger that will now get info messages
   */
  public void setLog(Logger newLog) {

    m_Log = newLog;
  }

  /**
   * Tells the panel to use a new set of instances.
   *
   * @param inst a set of Instances
   */
  public void setInstances(Instances inst) {

    m_Instances = inst;
  
    m_ignoreKeyModel.removeAllElements();
   
    String [] attribNames = new String [m_Instances.numAttributes()];
    for (int i = 0; i < m_Instances.numAttributes(); i++) {
      String name = m_Instances.attribute(i).name();
      m_ignoreKeyModel.addElement(name);
      String type = "(" + Attribute.typeToStringShort(m_Instances.attribute(i)) + ") ";
      String attnm = m_Instances.attribute(i).name();
      attribNames[i] = type + attnm;
    }

   
    m_StartBut.setEnabled(m_RunThread == null);
    m_StopBut.setEnabled(m_RunThread != null);
    m_ignoreBut.setEnabled(true);
    m_ClassCombo.setModel(new DefaultComboBoxModel(attribNames));
    if (inst.classIndex() == -1)
      m_ClassCombo.setSelectedIndex(attribNames.length - 1);
    else
      m_ClassCombo.setSelectedIndex(inst.classIndex());
    updateRadioLinks();
  }

  /**
   * Sets the user test set. Information about the current test set
   * is displayed in an InstanceSummaryPanel and the user is given the
   * ability to load another set from a file or url.
   *
   */
  protected void setTestSet() {

    if (m_SetTestFrame == null) {
      final SetInstancesPanel sp = new SetInstancesPanel();
      sp.setReadIncrementally(false);
      m_Summary = sp.getSummary();
      if (m_TestInstances != null) {
  sp.setInstances(m_TestInstances);
      }
      sp.addPropertyChangeListener(new PropertyChangeListener() {
  public void propertyChange(PropertyChangeEvent e) {
    m_TestInstances = sp.getInstances();
    m_TestInstances.setClassIndex(-1)// make sure that no class attribute is set!
  }
      });
      // Add propertychangelistener to update m_TestInstances whenever
      // it changes in the settestframe
      m_SetTestFrame = new JFrame("Test Instances");
      sp.setParentFrame(m_SetTestFrame);   // enable Close-Button
      m_SetTestFrame.getContentPane().setLayout(new BorderLayout());
      m_SetTestFrame.getContentPane().add(sp, BorderLayout.CENTER);
      m_SetTestFrame.pack();
    }
    m_SetTestFrame.setVisible(true);
  }
 
  /**
   * Starts running the currently configured clusterer with the current
   * settings. This is run in a separate thread, and will only start if
   * there is no clusterer already running. The clusterer output is sent
   * to the results history panel.
   */
  protected void startClusterer() {

    if (m_RunThread == null) {
      m_StartBut.setEnabled(false);
      m_StopBut.setEnabled(true);
      m_ignoreBut.setEnabled(false);
      m_RunThread = new Thread() {
  public void run() {
    // Copy the current state of things
    m_Log.statusMessage("Setting up...");
    Instances inst = new Instances(m_Instances);
    inst.setClassIndex(-1);
    Instances userTest = null;
    ClustererAssignmentsPlotInstances plotInstances = ExplorerDefaults.getClustererAssignmentsPlotInstances();
    plotInstances.setClusterer((Clusterer) m_ClustererEditor.getValue());
    if (m_TestInstances != null) {
      userTest = new Instances(m_TestInstances);
    }
   
    boolean saveVis = m_StorePredictionsBut.isSelected();
    String grph = null;
    int[] ignoredAtts = null;

    int testMode = 0;
    int percent = 66;
    Clusterer clusterer = (Clusterer) m_ClustererEditor.getValue();
    Clusterer fullClusterer = null;
    StringBuffer outBuff = new StringBuffer();
    String name = (new SimpleDateFormat("HH:mm:ss - ")).format(new Date());
    String cname = clusterer.getClass().getName();
    if (cname.startsWith("weka.clusterers.")) {
      name += cname.substring("weka.clusterers.".length());
    } else {
      name += cname;
    }
          String cmd = m_ClustererEditor.getValue().getClass().getName();
          if (m_ClustererEditor.getValue() instanceof OptionHandler)
            cmd += " " + Utils.joinOptions(((OptionHandler) m_ClustererEditor.getValue()).getOptions());
    try {
      m_Log.logMessage("Started " + cname);
      m_Log.logMessage("Command: " + cmd);
      if (m_Log instanceof TaskLogger) {
        ((TaskLogger)m_Log).taskStarted();
      }
      if (m_PercentBut.isSelected()) {
        testMode = 2;
        percent = Integer.parseInt(m_PercentText.getText());
        if ((percent <= 0) || (percent >= 100)) {
    throw new Exception("Percentage must be between 0 and 100");
        }
      } else if (m_TrainBut.isSelected()) {
        testMode = 3;
      } else if (m_TestSplitBut.isSelected()) {
        testMode = 4;
        // Check the test instance compatibility
        if (userTest == null) {
    throw new Exception("No user test set has been opened");
        }
        if (!inst.equalHeaders(userTest)) {
    throw new Exception("Train and test set are not compatible\n" + inst.equalHeadersMsg(userTest));
        }
      } else if (m_ClassesToClustersBut.isSelected()) {
        testMode = 5;
      } else {
        throw new Exception("Unknown test mode");
      }

      Instances trainInst = new Instances(inst);
      if (m_ClassesToClustersBut.isSelected()) {
        trainInst.setClassIndex(m_ClassCombo.getSelectedIndex());
        inst.setClassIndex(m_ClassCombo.getSelectedIndex());
        if (inst.classAttribute().isNumeric()) {
    throw new Exception("Class must be nominal for class based "
            +"evaluation!");
        }
      }
      if (!m_ignoreKeyList.isSelectionEmpty()) {
        trainInst = removeIgnoreCols(trainInst);
      }

      // Output some header information
      outBuff.append("=== Run information ===\n\n");
      outBuff.append("Scheme:       " + cname);
      if (clusterer instanceof OptionHandler) {
        String [] o = ((OptionHandler) clusterer).getOptions();
        outBuff.append(" " + Utils.joinOptions(o));
      }
      outBuff.append("\n");
      outBuff.append("Relation:     " + inst.relationName() + '\n');
      outBuff.append("Instances:    " + inst.numInstances() + '\n');
      outBuff.append("Attributes:   " + inst.numAttributes() + '\n');
      if (inst.numAttributes() < 100) {
        boolean [] selected = new boolean [inst.numAttributes()];
        for (int i = 0; i < inst.numAttributes(); i++) {
    selected[i] = true;
        }
        if (!m_ignoreKeyList.isSelectionEmpty()) {
    int [] indices = m_ignoreKeyList.getSelectedIndices();
    for (int i = 0; i < indices.length; i++) {
      selected[indices[i]] = false;
    }
        }
        if (m_ClassesToClustersBut.isSelected()) {
    selected[m_ClassCombo.getSelectedIndex()] = false;
        }
        for (int i = 0; i < inst.numAttributes(); i++) {
    if (selected[i]) {
      outBuff.append("              " + inst.attribute(i).name()
         + '\n');
    }
        }
        if (!m_ignoreKeyList.isSelectionEmpty()
      || m_ClassesToClustersBut.isSelected()) {
    outBuff.append("Ignored:\n");
    for (int i = 0; i < inst.numAttributes(); i++) {
      if (!selected[i]) {
        outBuff.append("              " + inst.attribute(i).name()
           + '\n');
      }
    }
        }
      } else {
        outBuff.append("              [list of attributes omitted]\n");
      }

      if (!m_ignoreKeyList.isSelectionEmpty()) {
        ignoredAtts = m_ignoreKeyList.getSelectedIndices();
      }

      if (m_ClassesToClustersBut.isSelected()) {
        // add class to ignored list
        if (ignoredAtts == null) {
    ignoredAtts = new int[1];
    ignoredAtts[0] = m_ClassCombo.getSelectedIndex();
        } else {
    int[] newIgnoredAtts = new int[ignoredAtts.length+1];
    System.arraycopy(ignoredAtts, 0, newIgnoredAtts, 0, ignoredAtts.length);
    newIgnoredAtts[ignoredAtts.length] = m_ClassCombo.getSelectedIndex();
    ignoredAtts = newIgnoredAtts;
        }
      }


      outBuff.append("Test mode:    ");
      switch (testMode) {
        case 3: // Test on training
        outBuff.append("evaluate on training data\n");
        break;
        case 2: // Percent split
        outBuff.append("split " + percent
             + "% train, remainder test\n");
        break;
        case 4: // Test on user split
        outBuff.append("user supplied test set: "
           + userTest.numInstances() + " instances\n");
        break;
      case 5: // Classes to clusters evaluation on training
        outBuff.append("Classes to clusters evaluation on training data");
       
        break;
      }
      outBuff.append("\n");
      m_History.addResult(name, outBuff);
      m_History.setSingle(name);
     
      // Build the model and output it.
      m_Log.statusMessage("Building model on training data...");

      // remove the class attribute (if set) and build the clusterer
      clusterer.buildClusterer(removeClass(trainInst));
     
      if (testMode == 2) {
        outBuff.append("\n=== Clustering model (full training set) ===\n\n");
     
        outBuff.append(clusterer.toString() + '\n');
      }
      m_History.updateResult(name);
      if (clusterer instanceof Drawable) {
        try {
    grph = ((Drawable)clusterer).graph();
        } catch (Exception ex) {
        }
      }
      // copy full model for output
      SerializedObject so = new SerializedObject(clusterer);
      fullClusterer = (Clusterer) so.getObject();
     
      ClusterEvaluation eval = new ClusterEvaluation();
      eval.setClusterer(clusterer);
      switch (testMode) {
        case 3: case 5: // Test on training
        m_Log.statusMessage("Clustering training data...");
        eval.evaluateClusterer(trainInst);
        plotInstances.setInstances(inst);
        plotInstances.setClusterEvaluation(eval);
        outBuff.append("=== Model and evaluation on training set ===\n\n");
        break;

        case 2: // Percent split
        m_Log.statusMessage("Randomizing instances...");
        inst.randomize(new Random(1));
        trainInst.randomize(new Random(1));
        int trainSize = trainInst.numInstances() * percent / 100;
        int testSize = trainInst.numInstances() - trainSize;
        Instances train = new Instances(trainInst, 0, trainSize);
        Instances test = new Instances(trainInst, trainSize, testSize);
        Instances testVis = new Instances(inst, trainSize, testSize);
        m_Log.statusMessage("Building model on training split...");
        clusterer.buildClusterer(train);
        m_Log.statusMessage("Evaluating on test split...");
        eval.evaluateClusterer(test);
        plotInstances.setInstances(testVis);
        plotInstances.setClusterEvaluation(eval);
        outBuff.append("=== Model and evaluation on test split ===\n");
        break;
   
        case 4: // Test on user split
        m_Log.statusMessage("Evaluating on test data...");
        Instances userTestT = new Instances(userTest);
        if (!m_ignoreKeyList.isSelectionEmpty()) {
    userTestT = removeIgnoreCols(userTestT);
        }
        eval.evaluateClusterer(userTestT);
        plotInstances.setInstances(userTest);
        plotInstances.setClusterEvaluation(eval);
        outBuff.append("=== Model and evaluation on test set ===\n");
        break;

        default:
        throw new Exception("Test mode not implemented");
      }
      outBuff.append(eval.clusterResultsToString());
      outBuff.append("\n");
      m_History.updateResult(name);
      m_Log.logMessage("Finished " + cname);
      m_Log.statusMessage("OK");
    } catch (Exception ex) {
      ex.printStackTrace();
      m_Log.logMessage(ex.getMessage());
      JOptionPane.showMessageDialog(ClustererPanel.this,
            "Problem evaluating clusterer:\n"
            + ex.getMessage(),
            "Evaluate clusterer",
            JOptionPane.ERROR_MESSAGE);
      m_Log.statusMessage("Problem evaluating clusterer");
    } finally {
      if (plotInstances != null) {
        plotInstances.setUp();
        m_CurrentVis = new VisualizePanel();
        m_CurrentVis.setName(name+" ("+inst.relationName()+")");
        m_CurrentVis.setLog(m_Log);
        try {
    m_CurrentVis.addPlot(plotInstances.getPlotData(name));
        } catch (Exception ex) {
    System.err.println(ex);
        }
        plotInstances.cleanUp();

        FastVector vv = new FastVector();
        vv.addElement(fullClusterer);
        Instances trainHeader = new Instances(m_Instances, 0);
        vv.addElement(trainHeader);
        if (ignoredAtts != null) vv.addElement(ignoredAtts);
        if (saveVis) {
    vv.addElement(m_CurrentVis);
    if (grph != null) {
      vv.addElement(grph);
    }
   
        }
        m_History.addObject(name, vv);
      }
      if (isInterrupted()) {
        m_Log.logMessage("Interrupted " + cname);
        m_Log.statusMessage("See error log");
      }
      m_RunThread = null;
      m_StartBut.setEnabled(true);
      m_StopBut.setEnabled(false);
      m_ignoreBut.setEnabled(true);
      if (m_Log instanceof TaskLogger) {
        ((TaskLogger)m_Log).taskFinished();
      }
    }
  }
      };
      m_RunThread.setPriority(Thread.MIN_PRIORITY);
      m_RunThread.start();
    }
  }

  private Instances removeClass(Instances inst) {
    Remove af = new Remove();
    Instances retI = null;
   
    try {
      if (inst.classIndex() < 0) {
  retI = inst;
      } else {
  af.setAttributeIndices(""+(inst.classIndex()+1));
  af.setInvertSelection(false);
  af.setInputFormat(inst);
  retI = Filter.useFilter(inst, af);
      }
    } catch (Exception e) {
      e.printStackTrace();
    }
    return retI;
  }

  private Instances removeIgnoreCols(Instances inst) {
   
    // If the user is doing classes to clusters evaluation and
    // they have opted to ignore the class, then unselect the class in
    // the ignore list
    if (m_ClassesToClustersBut.isSelected()) {
      int classIndex = m_ClassCombo.getSelectedIndex();
      if (m_ignoreKeyList.isSelectedIndex(classIndex)) {
  m_ignoreKeyList.removeSelectionInterval(classIndex, classIndex);
      }
    }
    int [] selected = m_ignoreKeyList.getSelectedIndices();
    Remove af = new Remove();
    Instances retI = null;

    try {
      af.setAttributeIndicesArray(selected);
      af.setInvertSelection(false);
      af.setInputFormat(inst);
      retI = Filter.useFilter(inst, af);
    } catch (Exception e) {
      e.printStackTrace();
    }
   
    return retI;
  }

  private Instances removeIgnoreCols(Instances inst, int[] toIgnore) {

    Remove af = new Remove();
    Instances retI = null;

    try {
      af.setAttributeIndicesArray(toIgnore);
      af.setInvertSelection(false);
      af.setInputFormat(inst);
      retI = Filter.useFilter(inst, af);
    } catch (Exception e) {
      e.printStackTrace();
    }
   
    return retI;
  }

  /**
   * Stops the currently running clusterer (if any).
   */
  protected void stopClusterer() {

    if (m_RunThread != null) {
      m_RunThread.interrupt();
     
      // This is deprecated (and theoretically the interrupt should do).
      m_RunThread.stop();
     
    }
  }

  /**
   * Pops up a TreeVisualizer for the clusterer from the currently
   * selected item in the results list
   * @param graphString the description of the tree in dotty format
   * @param treeName the title to assign to the display
   */
  protected void visualizeTree(String graphString, String treeName) {
    final javax.swing.JFrame jf =
      new javax.swing.JFrame("Weka Classifier Tree Visualizer: "+treeName);
    jf.setSize(500,400);
    jf.getContentPane().setLayout(new BorderLayout());
    if (graphString.contains("digraph")) {
      TreeVisualizer tv = new TreeVisualizer(null,
               graphString,
               new PlaceNode2());
      jf.getContentPane().add(tv, BorderLayout.CENTER);
      jf.addWindowListener(new java.awt.event.WindowAdapter() {
    public void windowClosing(java.awt.event.WindowEvent e) {
      jf.dispose();
    }
        });
      jf.setVisible(true);
      tv.fitToScreen();
    } else if (graphString.startsWith("Newick:")) {
      HierarchyVisualizer tv = new HierarchyVisualizer(graphString.substring(7));
      jf.getContentPane().add(tv, BorderLayout.CENTER);
      jf.addWindowListener(new java.awt.event.WindowAdapter() {
        public void windowClosing(java.awt.event.WindowEvent e) {
          jf.dispose();
        }
      });
      jf.setVisible(true);
      tv.fitToScreen();
    }
  }

  /**
   * Pops up a visualize panel to display cluster assignments
   * @param sp the visualize panel to display
   */
  protected void visualizeClusterAssignments(VisualizePanel sp) {
    if (sp != null) {
      String plotName = sp.getName();
      final javax.swing.JFrame jf =
  new javax.swing.JFrame("Weka Clusterer Visualize: "+plotName);
      jf.setSize(500,400);
      jf.getContentPane().setLayout(new BorderLayout());
      jf.getContentPane().add(sp, BorderLayout.CENTER);
      jf.addWindowListener(new java.awt.event.WindowAdapter() {
    public void windowClosing(java.awt.event.WindowEvent e) {
      jf.dispose();
    }
  });

      jf.setVisible(true);
    }
  }

  /**
   * Handles constructing a popup menu with visualization options
   * @param name the name of the result history list entry clicked on by
   * the user
   * @param x the x coordinate for popping up the menu
   * @param y the y coordinate for popping up the menu
   */
  protected void visualizeClusterer(String name, int x, int y) {
    final String selectedName = name;
    JPopupMenu resultListMenu = new JPopupMenu();

    JMenuItem visMainBuffer = new JMenuItem("View in main window");
    if (selectedName != null) {
      visMainBuffer.addActionListener(new ActionListener() {
    public void actionPerformed(ActionEvent e) {
      m_History.setSingle(selectedName);
    }
  });
    } else {
      visMainBuffer.setEnabled(false);
    }
    resultListMenu.add(visMainBuffer);

    JMenuItem visSepBuffer = new JMenuItem("View in separate window");
    if (selectedName != null) {
    visSepBuffer.addActionListener(new ActionListener() {
  public void actionPerformed(ActionEvent e) {
    m_History.openFrame(selectedName);
  }
      });
    } else {
      visSepBuffer.setEnabled(false);
    }
    resultListMenu.add(visSepBuffer);

    JMenuItem saveOutput = new JMenuItem("Save result buffer");
    if (selectedName != null) {
      saveOutput.addActionListener(new ActionListener() {
    public void actionPerformed(ActionEvent e) {
      saveBuffer(selectedName);
    }
  });
    } else {
      saveOutput.setEnabled(false);
    }
    resultListMenu.add(saveOutput);
   
    JMenuItem deleteOutput = new JMenuItem("Delete result buffer");
    if (selectedName != null) {
      deleteOutput.addActionListener(new ActionListener() {
  public void actionPerformed(ActionEvent e) {
    m_History.removeResult(selectedName);
  }
      });
    } else {
      deleteOutput.setEnabled(false);
    }
    resultListMenu.add(deleteOutput);

    resultListMenu.addSeparator();

    JMenuItem loadModel = new JMenuItem("Load model");
    loadModel.addActionListener(new ActionListener() {
  public void actionPerformed(ActionEvent e) {
    loadClusterer();
  }
      });
    resultListMenu.add(loadModel);

    FastVector o = null;
    if (selectedName != null) {
      o = (FastVector)m_History.getNamedObject(selectedName);
    }

    VisualizePanel temp_vp = null;
    String temp_grph = null;
    Clusterer temp_clusterer = null;
    Instances temp_trainHeader = null;
    int[] temp_ignoreAtts = null;
   
    if (o != null) {
      for (int i = 0; i < o.size(); i++) {
  Object temp = o.elementAt(i);
  if (temp instanceof Clusterer) {
    temp_clusterer = (Clusterer)temp;
  } else if (temp instanceof Instances) { // training header
    temp_trainHeader = (Instances)temp;
  } else if (temp instanceof int[]) { // ignored attributes
    temp_ignoreAtts = (int[])temp;
  } else if (temp instanceof VisualizePanel) { // normal errors
    temp_vp = (VisualizePanel)temp;
  } else if (temp instanceof String) { // graphable output
    temp_grph = (String)temp;
  }
      }
    }
     
    final VisualizePanel vp = temp_vp;
    final String grph = temp_grph;
    final Clusterer clusterer = temp_clusterer;
    final Instances trainHeader = temp_trainHeader;
    final int[] ignoreAtts = temp_ignoreAtts;
   
    JMenuItem saveModel = new JMenuItem("Save model");
    if (clusterer != null) {
      saveModel.addActionListener(new ActionListener() {
    public void actionPerformed(ActionEvent e) {
      saveClusterer(selectedName, clusterer, trainHeader, ignoreAtts);
    }
  });
    } else {
      saveModel.setEnabled(false);
    }
    resultListMenu.add(saveModel);
   
    JMenuItem reEvaluate =
      new JMenuItem("Re-evaluate model on current test set");
    if (clusterer != null && m_TestInstances != null) {
      reEvaluate.addActionListener(new ActionListener() {
    public void actionPerformed(ActionEvent e) {
      reevaluateModel(selectedName, clusterer, trainHeader, ignoreAtts);
    }
  });
    } else {
      reEvaluate.setEnabled(false);
    }
    resultListMenu.add(reEvaluate);
   
    resultListMenu.addSeparator();
   
    JMenuItem visClusts = new JMenuItem("Visualize cluster assignments");
    if (vp != null) {
      visClusts.addActionListener(new ActionListener() {
      public void actionPerformed(ActionEvent e) {
        visualizeClusterAssignments(vp);
      }
    });
     
    } else {
      visClusts.setEnabled(false);
    }
    resultListMenu.add(visClusts);

    JMenuItem visTree = new JMenuItem("Visualize tree");
    if (grph != null) {
      visTree.addActionListener(new ActionListener() {
    public void actionPerformed(ActionEvent e) {
      String title;
      if (vp != null) title = vp.getName();
      else title = selectedName;
      visualizeTree(grph, title);
    }
  });
    } else {
      visTree.setEnabled(false);
    }
    resultListMenu.add(visTree);

   
    // visualization plugins
    JMenu visPlugins = new JMenu("Plugins");
    boolean availablePlugins = false;
   
    // trees
    if (grph != null) {
      // trees
      Vector pluginsVector = GenericObjectEditor.getClassnames(TreeVisualizePlugin.class.getName());
      for (int i = 0; i < pluginsVector.size(); i++) {
  String className = (String) (pluginsVector.elementAt(i));
  try {
    TreeVisualizePlugin plugin = (TreeVisualizePlugin) Class.forName(className).newInstance();
    if (plugin == null)
      continue;
    availablePlugins = true;
    JMenuItem pluginMenuItem = plugin.getVisualizeMenuItem(grph, selectedName);
    Version version = new Version();
    if (pluginMenuItem != null) {
      if (version.compareTo(plugin.getMinVersion()) < 0)
        pluginMenuItem.setText(pluginMenuItem.getText() + " (weka outdated)");
      if (version.compareTo(plugin.getMaxVersion()) >= 0)
        pluginMenuItem.setText(pluginMenuItem.getText() + " (plugin outdated)");
      visPlugins.add(pluginMenuItem);
    }
  }
  catch (Exception e) {
    //e.printStackTrace();
  }
      }
    }

    if (availablePlugins)
      resultListMenu.add(visPlugins);
   
    resultListMenu.show(m_History.getList(), x, y);
  }
 
  /**
   * Save the currently selected clusterer output to a file.
   * @param name the name of the buffer to save
   */
  protected void saveBuffer(String name) {
    StringBuffer sb = m_History.getNamedBuffer(name);
    if (sb != null) {
      if (m_SaveOut.save(sb)) {
  m_Log.logMessage("Save successful.");
      }
    }
  }

  private void setIgnoreColumns() {
    ListSelectorDialog jd = new ListSelectorDialog(null, m_ignoreKeyList);

    // Open the dialog
    int result = jd.showDialog();
   
    if (result != ListSelectorDialog.APPROVE_OPTION) {
      // clear selected indices
      m_ignoreKeyList.clearSelection();
    }
    updateCapabilitiesFilter(m_ClustererEditor.getCapabilitiesFilter());
  }

  /**
   * Saves the currently selected clusterer
   */
  protected void saveClusterer(String name, Clusterer clusterer,
             Instances trainHeader, int[] ignoredAtts) {

    File sFile = null;
    boolean saveOK = true;

    int returnVal = m_FileChooser.showSaveDialog(this);
    if (returnVal == JFileChooser.APPROVE_OPTION) {
      sFile = m_FileChooser.getSelectedFile();
      if (!sFile.getName().toLowerCase().endsWith(MODEL_FILE_EXTENSION)) {
  sFile = new File(sFile.getParent(), sFile.getName()
       + MODEL_FILE_EXTENSION);
      }
      m_Log.statusMessage("Saving model to file...");
     
      try {
  OutputStream os = new FileOutputStream(sFile);
  if (sFile.getName().endsWith(".gz")) {
    os = new GZIPOutputStream(os);
  }
  ObjectOutputStream objectOutputStream = new ObjectOutputStream(os);
  objectOutputStream.writeObject(clusterer);
  if (trainHeader != null) objectOutputStream.writeObject(trainHeader);
  if (ignoredAtts != null) objectOutputStream.writeObject(ignoredAtts);
  objectOutputStream.flush();
  objectOutputStream.close();
      } catch (Exception e) {
 
  JOptionPane.showMessageDialog(null, e, "Save Failed",
              JOptionPane.ERROR_MESSAGE);
  saveOK = false;
      }
      if (saveOK)
  m_Log.logMessage("Saved model (" + name
       + ") to file '" + sFile.getName() + "'");
      m_Log.statusMessage("OK");
    }
  }

  /**
   * Loads a clusterer
   */
  protected void loadClusterer() {

    int returnVal = m_FileChooser.showOpenDialog(this);
    if (returnVal == JFileChooser.APPROVE_OPTION) {
      File selected = m_FileChooser.getSelectedFile();
      Clusterer clusterer = null;
      Instances trainHeader = null;
      int[] ignoredAtts = null;

      m_Log.statusMessage("Loading model from file...");

      try {
  InputStream is = new FileInputStream(selected);
  if (selected.getName().endsWith(".gz")) {
    is = new GZIPInputStream(is);
  }
  ObjectInputStream objectInputStream = new ObjectInputStream(is);
  clusterer = (Clusterer) objectInputStream.readObject();
  try { // see if we can load the header & ignored attribute info
    trainHeader = (Instances) objectInputStream.readObject();
    ignoredAtts = (int[]) objectInputStream.readObject();
  } catch (Exception e) {} // don't fuss if we can't
  objectInputStream.close();
      } catch (Exception e) {
 
  JOptionPane.showMessageDialog(null, e, "Load Failed",
              JOptionPane.ERROR_MESSAGE);
     

      m_Log.statusMessage("OK");
     
      if (clusterer != null) {
  m_Log.logMessage("Loaded model from file '" + selected.getName()+ "'");
  String name = (new SimpleDateFormat("HH:mm:ss - ")).format(new Date());
  String cname = clusterer.getClass().getName();
  if (cname.startsWith("weka.clusterers."))
    cname = cname.substring("weka.clusterers.".length());
  name += cname + " from file '" + selected.getName() + "'";
  StringBuffer outBuff = new StringBuffer();

  outBuff.append("=== Model information ===\n\n");
  outBuff.append("Filename:     " + selected.getName() + "\n");
  outBuff.append("Scheme:       " + clusterer.getClass().getName());
  if (clusterer instanceof OptionHandler) {
    String [] o = ((OptionHandler) clusterer).getOptions();
    outBuff.append(" " + Utils.joinOptions(o));
  }
  outBuff.append("\n");

  if (trainHeader != null) {

    outBuff.append("Relation:     " + trainHeader.relationName() + '\n');
    outBuff.append("Attributes:   " + trainHeader.numAttributes() + '\n');
    if (trainHeader.numAttributes() < 100) {
      boolean [] selectedAtts = new boolean [trainHeader.numAttributes()];
      for (int i = 0; i < trainHeader.numAttributes(); i++) {
        selectedAtts[i] = true;
      }
     
      if (ignoredAtts != null)
        for (int i=0; i<ignoredAtts.length; i++)
    selectedAtts[ignoredAtts[i]] = false;
    
      for (int i = 0; i < trainHeader.numAttributes(); i++) {
        if (selectedAtts[i]) {
    outBuff.append("              " + trainHeader.attribute(i).name()
             + '\n');
        }
      }
      if (ignoredAtts != null) {
        outBuff.append("Ignored:\n");
        for (int i=0; i<ignoredAtts.length; i++)
    outBuff.append("              "
             + trainHeader.attribute(ignoredAtts[i]).name()
             + '\n');
      }
    } else {
      outBuff.append("              [list of attributes omitted]\n");
    }
  } else {
    outBuff.append("\nTraining data unknown\n");
  }
 
  outBuff.append("\n=== Clustering model ===\n\n");
  outBuff.append(clusterer.toString() + "\n");

  m_History.addResult(name, outBuff);
  m_History.setSingle(name);
  FastVector vv = new FastVector();
  vv.addElement(clusterer);
  if (trainHeader != null) vv.addElement(trainHeader);
  if (ignoredAtts != null) vv.addElement(ignoredAtts);
  // allow visualization of graphable classifiers
  String grph = null;
  if (clusterer instanceof Drawable) {
    try {
      grph = ((Drawable)clusterer).graph();
    } catch (Exception ex) {
    }
  }
  if (grph != null) vv.addElement(grph);
 
  m_History.addObject(name, vv);
 
      }
    }
  }

  /**
   * Re-evaluates the named clusterer with the current test set. Unpredictable
   * things will happen if the data set is not compatible with the clusterer.
   *
   * @param name the name of the clusterer entry
   * @param clusterer the clusterer to evaluate
   * @param trainHeader the header of the training set
   * @param ignoredAtts ignored attributes
   */
  protected void reevaluateModel(final String name,
                                 final Clusterer clusterer,
         final Instances trainHeader,
                                 final int[] ignoredAtts) {

    if (m_RunThread == null) {
      m_StartBut.setEnabled(false);
      m_StopBut.setEnabled(true);
      m_ignoreBut.setEnabled(false);
      m_RunThread = new Thread() {
          public void run() {
            // Copy the current state of things
            m_Log.statusMessage("Setting up...");

            StringBuffer outBuff = m_History.getNamedBuffer(name);
            Instances userTest = null;

            ClustererAssignmentsPlotInstances plotInstances = ExplorerDefaults.getClustererAssignmentsPlotInstances();
            plotInstances.setClusterer(clusterer);
            if (m_TestInstances != null) {
              userTest = new Instances(m_TestInstances);
            }
   
            boolean saveVis = m_StorePredictionsBut.isSelected();
            String grph = null;

            try {
              if (userTest == null) {
                throw new Exception("No user test set has been opened");
              }
              if (trainHeader != null && !trainHeader.equalHeaders(userTest)) {
                throw new Exception("Train and test set are not compatible\n" + trainHeader.equalHeadersMsg(userTest));
              }

              m_Log.statusMessage("Evaluating on test data...");
              m_Log.logMessage("Re-evaluating clusterer (" + name + ") on test set");

              m_Log.logMessage("Started reevaluate model");
              if (m_Log instanceof TaskLogger) {
                ((TaskLogger)m_Log).taskStarted();
              }
              ClusterEvaluation eval = new ClusterEvaluation();
              eval.setClusterer(clusterer);
   
              Instances userTestT = new Instances(userTest);
              if (ignoredAtts != null) {
                userTestT = removeIgnoreCols(userTestT, ignoredAtts);
              }

              eval.evaluateClusterer(userTestT);
     
              plotInstances.setClusterEvaluation(eval);
              plotInstances.setInstances(userTest);
              plotInstances.setUp();

              outBuff.append("\n=== Re-evaluation on test set ===\n\n");
              outBuff.append("User supplied test set\n")
              outBuff.append("Relation:     " + userTest.relationName() + '\n');
              outBuff.append("Instances:    " + userTest.numInstances() + '\n');
              outBuff.append("Attributes:   " + userTest.numAttributes() + "\n\n");
              if (trainHeader == null)
                outBuff.append("NOTE - if test set is not compatible then results are "
                               + "unpredictable\n\n");
     
              outBuff.append(eval.clusterResultsToString());
              outBuff.append("\n");
              m_History.updateResult(name);
              m_Log.logMessage("Finished re-evaluation");
              m_Log.statusMessage("OK");
            } catch (Exception ex) {
              ex.printStackTrace();
              m_Log.logMessage(ex.getMessage());
              JOptionPane.showMessageDialog(ClustererPanel.this,
                                            "Problem evaluating clusterer:\n"
                                            + ex.getMessage(),
                                            "Evaluate clusterer",
                                            JOptionPane.ERROR_MESSAGE);
              m_Log.statusMessage("Problem evaluating clusterer");

            } finally {
              if (plotInstances != null) {
                m_CurrentVis = new VisualizePanel();
                m_CurrentVis.setName(name+" ("+userTest.relationName()+")");
                m_CurrentVis.setLog(m_Log);
                try {
                  m_CurrentVis.addPlot(plotInstances.getPlotData(name));
                } catch (Exception ex) {
                  System.err.println(ex);
                }
 
                FastVector vv = new FastVector();
                vv.addElement(clusterer);
                if (trainHeader != null) vv.addElement(trainHeader);
                if (ignoredAtts != null) vv.addElement(ignoredAtts);
                if (saveVis) {
                  vv.addElement(m_CurrentVis);
                  if (grph != null) {
                    vv.addElement(grph);
                  }
   
                }
                m_History.addObject(name, vv);

              }
              if (isInterrupted()) {
                m_Log.logMessage("Interrupted reevaluate model");
                m_Log.statusMessage("See error log");
              }
              m_RunThread = null;
              m_StartBut.setEnabled(true);
              m_StopBut.setEnabled(false);
              m_ignoreBut.setEnabled(true);
              if (m_Log instanceof TaskLogger) {
                ((TaskLogger)m_Log).taskFinished();
              }
            }
          }
     
        };
      m_RunThread.setPriority(Thread.MIN_PRIORITY);
      m_RunThread.start();
    }
  }
 
  /**
   * updates the capabilities filter of the GOE
   *
   * @param filter  the new filter to use
   */
  protected void updateCapabilitiesFilter(Capabilities filter) {
    Instances     tempInst;
    Capabilities   filterClass;

    if (filter == null) {
      m_ClustererEditor.setCapabilitiesFilter(new Capabilities(null));
      return;
    }
   
    if (!ExplorerDefaults.getInitGenericObjectEditorFilter())
      tempInst = new Instances(m_Instances, 0);
    else
      tempInst = new Instances(m_Instances);
    tempInst.setClassIndex(-1);
   
    if (!m_ignoreKeyList.isSelectionEmpty()) {
      tempInst = removeIgnoreCols(tempInst);
    }

    try {
      filterClass = Capabilities.forInstances(tempInst);
    }
    catch (Exception e) {
      filterClass = new Capabilities(null);
    }
   
    m_ClustererEditor.setCapabilitiesFilter(filterClass);
   
    // check capabilities
    m_StartBut.setEnabled(true);
    Capabilities currentFilter = m_ClustererEditor.getCapabilitiesFilter();
    Clusterer clusterer = (Clusterer) m_ClustererEditor.getValue();
    Capabilities currentSchemeCapabilities =  null;
    if (clusterer != null && currentFilter != null &&
        (clusterer instanceof CapabilitiesHandler)) {
      currentSchemeCapabilities = ((CapabilitiesHandler)clusterer).getCapabilities();
     
      if (!currentSchemeCapabilities.supportsMaybe(currentFilter) &&
          !currentSchemeCapabilities.supports(currentFilter)) {
        m_StartBut.setEnabled(false);
      }
    }
  }
 
  /**
   * method gets called in case of a change event
   *
   * @param e    the associated change event
   */
  public void capabilitiesFilterChanged(CapabilitiesFilterChangeEvent e) {
    if (e.getFilter() == null)
      updateCapabilitiesFilter(null);
    else
      updateCapabilitiesFilter((Capabilities) e.getFilter().clone());
  }

  /**
   * Sets the Explorer to use as parent frame (used for sending notifications
   * about changes in the data)
   *
   * @param parent  the parent frame
   */
  public void setExplorer(Explorer parent) {
    m_Explorer = parent;
  }
 
  /**
   * returns the parent Explorer frame
   *
   * @return    the parent
   */
  public Explorer getExplorer() {
    return m_Explorer;
  }
 
  /**
   * Returns the title for the tab in the Explorer
   *
   * @return     the title of this tab
   */
  public String getTabTitle() {
    return "Cluster";
  }
 
  /**
   * Returns the tooltip for the tab in the Explorer
   *
   * @return     the tooltip of this tab
   */
  public String getTabTitleToolTip() {
    return "Identify instance clusters";
  }

  /**
   * Tests out the clusterer panel from the command line.
   *
   * @param args may optionally contain the name of a dataset to load.
   */
  public static void main(String [] args) {

    try {
      final javax.swing.JFrame jf =
  new javax.swing.JFrame("Weka Explorer: Cluster");
      jf.getContentPane().setLayout(new BorderLayout());
      final ClustererPanel sp = new ClustererPanel();
      jf.getContentPane().add(sp, BorderLayout.CENTER);
      weka.gui.LogPanel lp = new weka.gui.LogPanel();
      sp.setLog(lp);
      jf.getContentPane().add(lp, BorderLayout.SOUTH);
      jf.addWindowListener(new java.awt.event.WindowAdapter() {
  public void windowClosing(java.awt.event.WindowEvent e) {
    jf.dispose();
    System.exit(0);
  }
      });
      jf.pack();
      jf.setSize(800, 600);
      jf.setVisible(true);
      if (args.length == 1) {
  System.err.println("Loading instances from " + args[0]);
  java.io.Reader r = new java.io.BufferedReader(
         new java.io.FileReader(args[0]));
  Instances i = new Instances(r);
  sp.setInstances(i);
      }
    } catch (Exception ex) {
      ex.printStackTrace();
      System.err.println(ex.getMessage());
    }
  }
}
TOP

Related Classes of weka.gui.explorer.ClustererPanel

TOP
Copyright © 2015 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.