Package weka.gui.beans

Source Code of weka.gui.beans.ClassifierPerformanceEvaluator

/*
*    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.
*/

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

package weka.gui.beans;

import weka.classifiers.Classifier;
import weka.classifiers.AbstractClassifier;
import weka.classifiers.Evaluation;
import weka.classifiers.evaluation.ThresholdCurve;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.OptionHandler;
import weka.core.Utils;
import weka.gui.explorer.ClassifierErrorsPlotInstances;
import weka.gui.explorer.ExplorerDefaults;
import weka.gui.visualize.PlotData2D;

import java.io.Serializable;
import java.util.Enumeration;
import java.util.Vector;

/**
* A bean that evaluates the performance of batch trained classifiers
*
* @author <a href="mailto:mhall@cs.waikato.ac.nz">Mark Hall</a>
* @version $Revision: 6804 $
*/
public class ClassifierPerformanceEvaluator
  extends AbstractEvaluator
  implements BatchClassifierListener,
       Serializable, UserRequestAcceptor, EventConstraints {

  /** for serialization */
  private static final long serialVersionUID = -3511801418192148690L;

  /**
   * Evaluation object used for evaluating a classifier
   */
  private transient Evaluation m_eval;

  private transient Thread m_evaluateThread = null;
 
  private transient long m_currentBatchIdentifier;
  private transient int m_setsComplete;
 
  private Vector m_textListeners = new Vector();
  private Vector m_thresholdListeners = new Vector();
  private Vector m_visualizableErrorListeners = new Vector();

  public ClassifierPerformanceEvaluator() {
    m_visual.loadIcons(BeanVisual.ICON_PATH
           +"ClassifierPerformanceEvaluator.gif",
           BeanVisual.ICON_PATH
           +"ClassifierPerformanceEvaluator_animated.gif");
    m_visual.setText("ClassifierPerformanceEvaluator");
  }

  /**
   * Set a custom (descriptive) name for this bean
   *
   * @param name the name to use
   */
  public void setCustomName(String name) {
    m_visual.setText(name);
  }

  /**
   * Get the custom (descriptive) name for this bean (if one has been set)
   *
   * @return the custom name (or the default name)
   */
  public String getCustomName() {
    return m_visual.getText();
  }
 
  /**
   * Global info for this bean
   *
   * @return a <code>String</code> value
   */
  public String globalInfo() {
    return "Evaluate the performance of batch trained classifiers.";
  }

  // ----- Stuff for ROC curves
  private boolean m_rocListenersConnected = false;
 
  /** for generating plottable instance with predictions appended. */
  private transient ClassifierErrorsPlotInstances m_PlotInstances = null;
 
  protected static Evaluation adjustForInputMappedClassifier(Evaluation eval,
      weka.classifiers.Classifier classifier,
      Instances inst, ClassifierErrorsPlotInstances plotInstances) throws Exception {

    if (classifier instanceof weka.classifiers.misc.InputMappedClassifier) {
      Instances mappedClassifierHeader =
        ((weka.classifiers.misc.InputMappedClassifier)classifier).
        getModelHeader(new Instances(inst, 0));

      eval = new Evaluation(new Instances(mappedClassifierHeader, 0));

      if (!eval.getHeader().equalHeaders(inst)) {
        // When the InputMappedClassifier is loading a model,
        // we need to make a new dataset that maps the test instances to
        // the structure expected by the mapped classifier - this is only
        // to ensure that the ClassifierPlotInstances object is configured
        // in accordance with what the embeded classifier was trained with
        Instances mappedClassifierDataset =
          ((weka.classifiers.misc.InputMappedClassifier)classifier).
          getModelHeader(new Instances(mappedClassifierHeader, 0));
        for (int zz = 0; zz < inst.numInstances(); zz++) {
          Instance mapped = ((weka.classifiers.misc.InputMappedClassifier)classifier).
          constructMappedInstance(inst.instance(zz));
          mappedClassifierDataset.add(mapped);
        }
       
        eval.setPriors(mappedClassifierDataset);
        plotInstances.setInstances(mappedClassifierDataset);
        plotInstances.setClassifier(classifier);
        plotInstances.setClassIndex(mappedClassifierDataset.classIndex());
        plotInstances.setEvaluation(eval);
      }
    }
   
    return eval;
  }

  /**
   * Accept a classifier to be evaluated.
   *
   * @param ce a <code>BatchClassifierEvent</code> value
   */
  public void acceptClassifier(final BatchClassifierEvent ce) {
    if (ce.getTestSet() == null || ce.getTestSet().isStructureOnly()) {
      return; // cant evaluate empty/non-existent test instances
    }
    try {
      if (m_evaluateThread == null) {
  m_evaluateThread = new Thread() {
      public void run() {
        boolean errorOccurred = false;
//        final String oldText = m_visual.getText();
        Classifier classifier = ce.getClassifier();
        try {
    // if (ce.getSetNumber() == 1) {
          if (ce.getGroupIdentifier() != m_currentBatchIdentifier) {
     
      if (ce.getTrainSet().getDataSet() == null ||
          ce.getTrainSet().getDataSet().numInstances() == 0) {
        // we have no training set to estimate majority class
        // or mean of target from
        m_eval = new Evaluation(ce.getTestSet().getDataSet());
        m_PlotInstances = ExplorerDefaults.getClassifierErrorsPlotInstances();
        m_PlotInstances.setInstances(ce.getTestSet().getDataSet());
        m_PlotInstances.setClassifier(ce.getClassifier());
        m_PlotInstances.setClassIndex(ce.getTestSet().getDataSet().classIndex());
        m_PlotInstances.setEvaluation(m_eval);

        m_eval = adjustForInputMappedClassifier(m_eval, ce.getClassifier(),
            ce.getTestSet().getDataSet(), m_PlotInstances);
        m_eval.useNoPriors();
      } else {
        // we can set up with the training set here
        m_eval = new Evaluation(ce.getTrainSet().getDataSet());
        m_PlotInstances = ExplorerDefaults.getClassifierErrorsPlotInstances();
        m_PlotInstances.setInstances(ce.getTrainSet().getDataSet());
        m_PlotInstances.setClassifier(ce.getClassifier());
        m_PlotInstances.setClassIndex(ce.getTestSet().getDataSet().classIndex());
        m_PlotInstances.setEvaluation(m_eval);
       
        m_eval = adjustForInputMappedClassifier(m_eval, ce.getClassifier(),
                        ce.getTrainSet().getDataSet(), m_PlotInstances);
      }
//      m_classifier = ce.getClassifier();

      m_PlotInstances.setUp();
     
      m_currentBatchIdentifier = ce.getGroupIdentifier();
      m_setsComplete = 0;
    }
//    if (ce.getSetNumber() <= ce.getMaxSetNumber()) {
          if (m_setsComplete < ce.getMaxSetNumber()) {
     
      /*if (ce.getTrainSet().getDataSet() != null &&
          ce.getTrainSet().getDataSet().numInstances() > 0) {
        // set the priors
        m_eval.setPriors(ce.getTrainSet().getDataSet());
      } */
     
//      m_visual.setText("Evaluating ("+ce.getSetNumber()+")...");
      if (m_logger != null) {
        m_logger.statusMessage(statusMessagePrefix()
             +"Evaluating ("+ce.getSetNumber()
             +")...");
      }
      m_visual.setAnimated();
      /*
      m_eval.evaluateModel(ce.getClassifier(),
      ce.getTestSet().getDataSet()); */
      for (int i = 0; i < ce.getTestSet().getDataSet().numInstances(); i++) {
        Instance temp = ce.getTestSet().getDataSet().instance(i);
        m_PlotInstances.process(temp, ce.getClassifier(), m_eval);
      }
     
      m_setsComplete++;
    }
   
//    if (ce.getSetNumber() == ce.getMaxSetNumber()) {
          if (m_setsComplete == ce.getMaxSetNumber()) {
                  //      System.err.println(m_eval.toSummaryString());
      // m_resultsString.append(m_eval.toSummaryString());
      // m_outText.setText(m_resultsString.toString());
      String textTitle = classifier.getClass().getName();
      String textOptions = "";
      if (classifier instanceof OptionHandler) {
               textOptions =
                 Utils.joinOptions(((OptionHandler)classifier).getOptions());
      }
      textTitle =
        textTitle.substring(textTitle.lastIndexOf('.')+1,
          textTitle.length());
      String resultT = "=== Evaluation result ===\n\n"
        + "Scheme: " + textTitle + "\n"
        + ((textOptions.length() > 0) ? "Options: " + textOptions + "\n": "")
        + "Relation: " + ce.getTestSet().getDataSet().relationName()
        + "\n\n" + m_eval.toSummaryString();
                 
                  if (ce.getTestSet().getDataSet().
                      classAttribute().isNominal()) {
                    resultT += "\n" + m_eval.toClassDetailsString()
                      + "\n" + m_eval.toMatrixString();
                  }
                 
      TextEvent te =
        new TextEvent(ClassifierPerformanceEvaluator.this,
          resultT,
          textTitle);
      notifyTextListeners(te);

                  // set up visualizable errors
                  if (m_visualizableErrorListeners.size() > 0) {
                    PlotData2D errorD = m_PlotInstances.getPlotData(
                  textTitle + " " + textOptions);
                    VisualizableErrorEvent vel =
                      new VisualizableErrorEvent(ClassifierPerformanceEvaluator.this, errorD);
                    notifyVisualizableErrorListeners(vel);
                    m_PlotInstances.cleanUp();
                  }
                 

      if (ce.getTestSet().getDataSet().classAttribute().isNominal() &&
          m_thresholdListeners.size() > 0) {
        ThresholdCurve tc = new ThresholdCurve();
        Instances result = tc.getCurve(m_eval.predictions(), 0);
        result.
          setRelationName(ce.getTestSet().getDataSet().relationName());
        PlotData2D pd = new PlotData2D(result);
        String htmlTitle = "<html><font size=-2>"
          + textTitle;
        String newOptions = "";
        if (classifier instanceof OptionHandler) {
          String[] options =
            ((OptionHandler) classifier).getOptions();
          if (options.length > 0) {
            for (int ii = 0; ii < options.length; ii++) {
              if (options[ii].length() == 0) {
                continue;
              }
              if (options[ii].charAt(0) == '-' &&
                  !(options[ii].charAt(1) >= '0' &&
                      options[ii].charAt(1)<= '9')) {
                newOptions += "<br>";
              }
              newOptions += options[ii];
            }
          }
        }
       
       htmlTitle += " " + newOptions + "<br>"
          + " (class: "
                      +ce.getTestSet().getDataSet().
                        classAttribute().value(0) + ")"
                      + "</font></html>";
        pd.setPlotName(textTitle + " (class: "
                        +ce.getTestSet().getDataSet().
                          classAttribute().value(0) + ")");
        pd.setPlotNameHTML(htmlTitle);
        boolean [] connectPoints =
          new boolean [result.numInstances()];
        for (int jj = 1; jj < connectPoints.length; jj++) {
          connectPoints[jj] = true;
        }
        pd.setConnectPoints(connectPoints);
        ThresholdDataEvent rde =
          new ThresholdDataEvent(ClassifierPerformanceEvaluator.this,
               pd, ce.getTestSet().getDataSet().classAttribute());
        notifyThresholdListeners(rde);
        /*te = new TextEvent(ClassifierPerformanceEvaluator.this,
               result.toString(),
               "ThresholdCurveInst");
               notifyTextListeners(te); */
      }
      if (m_logger != null) {
        m_logger.statusMessage(statusMessagePrefix() + "Finished.");
      }

      // save memory
      m_PlotInstances = null;
    }
        } catch (Exception ex) {
          errorOccurred = true;
          ClassifierPerformanceEvaluator.this.stop(); // stop all processing
          if (m_logger != null) {
            m_logger.logMessage("[ClassifierPerformanceEvaluator] "
                + statusMessagePrefix()
                + " problem evaluating classifier. "
                + ex.getMessage());
          }
    ex.printStackTrace();
        } finally {
//    m_visual.setText(oldText);
    m_visual.setStatic();
    m_evaluateThread = null;
           
    if (m_logger != null) {
      if (errorOccurred) {
        m_logger.statusMessage(statusMessagePrefix()
            + "ERROR (See log for details)");
      } else if (isInterrupted()) {
        m_logger.logMessage("[" + getCustomName() +"] Evaluation interrupted!");
        m_logger.statusMessage(statusMessagePrefix()
            + "INTERRUPTED");
      }
    }
    block(false);
        }
      }
    };
  m_evaluateThread.setPriority(Thread.MIN_PRIORITY);
  m_evaluateThread.start();

  // make sure the thread is still running before we block
  //  if (m_evaluateThread.isAlive()) {
  block(true);
    //  }
  m_evaluateThread = null;
      }
    }  catch (Exception ex) {
      ex.printStackTrace();
    }
  }

  /**
   * Returns true if. at this time, the bean is busy with some
   * (i.e. perhaps a worker thread is performing some calculation).
   *
   * @return true if the bean is busy.
   */
  public boolean isBusy() {
    return (m_evaluateThread != null);
  }
   
  /**
   * Try and stop any action
   */
  public void stop() {
    // tell the listenee (upstream bean) to stop
    if (m_listenee instanceof BeanCommon) {
      //      System.err.println("Listener is BeanCommon");
      ((BeanCommon)m_listenee).stop();
    }

    // stop the evaluate thread
    if (m_evaluateThread != null) {
      m_evaluateThread.interrupt();
      m_evaluateThread.stop();
      m_evaluateThread = null;
      m_visual.setStatic();
    }
  }
 
  /**
   * Function used to stop code that calls acceptClassifier. This is
   * needed as classifier evaluation is performed inside a separate
   * thread of execution.
   *
   * @param tf a <code>boolean</code> value
   */
  private synchronized void block(boolean tf) {
    if (tf) {
      try {
  // only block if thread is still doing something useful!
  if (m_evaluateThread != null && m_evaluateThread.isAlive()) {
    wait();
  }
      } catch (InterruptedException ex) {
      }
    } else {
      notifyAll();
    }
  }

  /**
   * Return an enumeration of user activated requests for this bean
   *
   * @return an <code>Enumeration</code> value
   */
  public Enumeration enumerateRequests() {
    Vector newVector = new Vector(0);
    if (m_evaluateThread != null) {
      newVector.addElement("Stop");
    }
    return newVector.elements();
  }

  /**
   * Perform the named request
   *
   * @param request the request to perform
   * @exception IllegalArgumentException if an error occurs
   */
  public void performRequest(String request) {
    if (request.compareTo("Stop") == 0) {
      stop();
    } else {
      throw new
  IllegalArgumentException(request

        + " not supported (ClassifierPerformanceEvaluator)");
    }
  }

  /**
   * Add a text listener
   *
   * @param cl a <code>TextListener</code> value
   */
  public synchronized void addTextListener(TextListener cl) {
    m_textListeners.addElement(cl);
  }

  /**
   * Remove a text listener
   *
   * @param cl a <code>TextListener</code> value
   */
  public synchronized void removeTextListener(TextListener cl) {
    m_textListeners.remove(cl);
  }
 
  /**
   * Add a threshold data listener
   *
   * @param cl a <code>ThresholdDataListener</code> value
   */
  public synchronized void addThresholdDataListener(ThresholdDataListener cl) {
    m_thresholdListeners.addElement(cl);
  }

  /**
   * Remove a Threshold data listener
   *
   * @param cl a <code>ThresholdDataListener</code> value
   */
  public synchronized void removeThresholdDataListener(ThresholdDataListener cl) {
    m_thresholdListeners.remove(cl);
  }

  /**
   * Add a visualizable error listener
   *
   * @param vel a <code>VisualizableErrorListener</code> value
   */
  public synchronized void addVisualizableErrorListener(VisualizableErrorListener vel) {
    m_visualizableErrorListeners.add(vel);
  }

  /**
   * Remove a visualizable error listener
   *
   * @param vel a <code>VisualizableErrorListener</code> value
   */
  public synchronized void removeVisualizableErrorListener(VisualizableErrorListener vel) {
    m_visualizableErrorListeners.remove(vel);
  }

  /**
   * Notify all text listeners of a TextEvent
   *
   * @param te a <code>TextEvent</code> value
   */
  private void notifyTextListeners(TextEvent te) {
    Vector l;
    synchronized (this) {
      l = (Vector)m_textListeners.clone();
    }
    if (l.size() > 0) {
      for(int i = 0; i < l.size(); i++) {
  //  System.err.println("Notifying text listeners "
  //         +"(ClassifierPerformanceEvaluator)");
  ((TextListener)l.elementAt(i)).acceptText(te);
      }
    }
  }

  /**
   * Notify all ThresholdDataListeners of a ThresholdDataEvent
   *
   * @param te a <code>ThresholdDataEvent</code> value
   */
  private void notifyThresholdListeners(ThresholdDataEvent re) {
    Vector l;
    synchronized (this) {
      l = (Vector)m_thresholdListeners.clone();
    }
    if (l.size() > 0) {
      for(int i = 0; i < l.size(); i++) {
  //  System.err.println("Notifying text listeners "
  //         +"(ClassifierPerformanceEvaluator)");
  ((ThresholdDataListener)l.elementAt(i)).acceptDataSet(re);
      }
    }
  }

  /**
   * Notify all VisualizableErrorListeners of a VisualizableErrorEvent
   *
   * @param te a <code>VisualizableErrorEvent</code> value
   */
  private void notifyVisualizableErrorListeners(VisualizableErrorEvent re) {
    Vector l;
    synchronized (this) {
      l = (Vector)m_visualizableErrorListeners.clone();
    }
    if (l.size() > 0) {
      for(int i = 0; i < l.size(); i++) {
  //  System.err.println("Notifying text listeners "
  //         +"(ClassifierPerformanceEvaluator)");
  ((VisualizableErrorListener)l.elementAt(i)).acceptDataSet(re);
      }
    }
  }

  /**
   * Returns true, if at the current time, the named event could
   * be generated. Assumes that supplied event names are names of
   * events that could be generated by this bean.
   *
   * @param eventName the name of the event in question
   * @return true if the named event could be generated at this point in
   * time
   */
  public boolean eventGeneratable(String eventName) {
    if (m_listenee == null) {
      return false;
    }

    if (m_listenee instanceof EventConstraints) {
      if (!((EventConstraints)m_listenee).
    eventGeneratable("batchClassifier")) {
  return false;
      }
    }
    return true;
  }
 
  private String statusMessagePrefix() {
    return getCustomName() + "$" + hashCode() + "|";
  }
}
TOP

Related Classes of weka.gui.beans.ClassifierPerformanceEvaluator

TOP
Copyright © 2018 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.