experiment.java

Java 编写的多种数据挖掘算法 包括聚类、分类、预处理等

/*
 *    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.
 */

/*
 *    Experiment.java
 *    Copyright (C) 1999 Len Trigg
 *
 */


package weka.experiment;

import weka.core.AdditionalMeasureProducer;
import weka.core.FastVector;
import weka.core.Instances;
import weka.core.Option;
import weka.core.OptionHandler;
import weka.core.Utils;
import weka.core.xml.KOML;
import weka.core.xml.XMLOptions;
import weka.experiment.xml.XMLExperiment;

import java.beans.PropertyDescriptor;
import java.io.BufferedInputStream;
import java.io.BufferedOutputStream;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.io.FileReader;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.io.Reader;
import java.io.Serializable;
import java.lang.reflect.Array;
import java.lang.reflect.Method;
import java.util.Enumeration;
import java.util.Vector;

import javax.swing.DefaultListModel;

/**
 * Holds all the necessary configuration information for a standard
 * type experiment. This object is able to be serialized for storage
 * on disk.
 *
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -L &lt;num&gt;
 *  The lower run number to start the experiment from.
 *  (default 1)</pre>
 * 
 * <pre> -U &lt;num&gt;
 *  The upper run number to end the experiment at (inclusive).
 *  (default 10)</pre>
 * 
 * <pre> -T &lt;arff file&gt;
 *  The dataset to run the experiment on.
 *  (required, may be specified multiple times)</pre>
 * 
 * <pre> -P &lt;class name&gt;
 *  The full class name of a ResultProducer (required).
 *  eg: weka.experiment.RandomSplitResultProducer</pre>
 * 
 * <pre> -D &lt;class name&gt;
 *  The full class name of a ResultListener (required).
 *  eg: weka.experiment.CSVResultListener</pre>
 * 
 * <pre> -N &lt;string&gt;
 *  A string containing any notes about the experiment.
 *  (default none)</pre>
 * 
 * <pre> 
 * Options specific to result producer weka.experiment.RandomSplitResultProducer:
 * </pre>
 * 
 * <pre> -P &lt;percent&gt;
 *  The percentage of instances to use for training.
 *  (default 66)</pre>
 * 
 * <pre> -D
 * Save raw split evaluator output.</pre>
 * 
 * <pre> -O &lt;file/directory name/path&gt;
 *  The filename where raw output will be stored.
 *  If a directory name is specified then then individual
 *  outputs will be gzipped, otherwise all output will be
 *  zipped to the named file. Use in conjuction with -D. (default splitEvalutorOut.zip)</pre>
 * 
 * <pre> -W &lt;class name&gt;
 *  The full class name of a SplitEvaluator.
 *  eg: weka.experiment.ClassifierSplitEvaluator</pre>
 * 
 * <pre> -R
 *  Set when data is not to be randomized and the data sets' size.
 *  Is not to be determined via probabilistic rounding.</pre>
 * 
 * <pre> 
 * Options specific to split evaluator weka.experiment.ClassifierSplitEvaluator:
 * </pre>
 * 
 * <pre> -W &lt;class name&gt;
 *  The full class name of the classifier.
 *  eg: weka.classifiers.bayes.NaiveBayes</pre>
 * 
 * <pre> -C &lt;index&gt;
 *  The index of the class for which IR statistics
 *  are to be output. (default 1)</pre>
 * 
 * <pre> -I &lt;index&gt;
 *  The index of an attribute to output in the
 *  results. This attribute should identify an
 *  instance in order to know which instances are
 *  in the test set of a cross validation. if 0
 *  no output (default 0).</pre>
 * 
 * <pre> -P
 *  Add target and prediction columns to the result
 *  for each fold.</pre>
 * 
 * <pre> 
 * Options specific to classifier weka.classifiers.rules.ZeroR:
 * </pre>
 * 
 * <pre> -D
 *  If set, classifier is run in debug mode and
 *  may output additional info to the console</pre>
 * 
 <!-- options-end -->
 *
 * All options after -- will be passed to the result producer. <p>
 *
 * @author Len Trigg ([e-mail])
 * @version $Revision: 1.24 $
 */
public class Experiment 
  implements Serializable, OptionHandler {
  
  /** for serialization */
  static final long serialVersionUID = 44945596742646663L;
  
  /** The filename extension that should be used for experiment files */
  public static String FILE_EXTENSION = ".exp";

  /** Where results will be sent */
  protected ResultListener m_ResultListener = new InstancesResultListener();
  
  /** The result producer */
  protected ResultProducer m_ResultProducer = new RandomSplitResultProducer();

  /** Lower run number */
  protected int m_RunLower = 1;

  /** Upper run number */
  protected int m_RunUpper = 10;

  /** An array of dataset files */
  protected DefaultListModel m_Datasets = new DefaultListModel();

  /** True if the exp should also iterate over a property of the RP */
  protected boolean m_UsePropertyIterator = false;
  
  /** The path to the iterator property */
  protected PropertyNode [] m_PropertyPath;
  
  /** The array of values to set the property to */
  protected Object m_PropertyArray;

  /** User notes about the experiment */
  protected String m_Notes = "";

  /** Method names of additional measures of objects contained in the 
      custom property iterator. Only methods names beginning with "measure"
      and returning doubles are recognised */
  protected String [] m_AdditionalMeasures = null;

  /** True if the class attribute is the first attribute for all
      datasets involved in this experiment. */
  protected boolean m_ClassFirst = false;

  /** If true an experiment will advance the current data set befor
      any custom itererator */
  protected boolean m_AdvanceDataSetFirst = true;

  /**
   * Sets whether the first attribute is treated as the class
   * for all datasets involved in the experiment. This information
   * is not output with the result of the experiments!
   * 
   * @param flag	whether the class attribute is the first and not the last
   */
  public void classFirst(boolean flag) {
    
    m_ClassFirst = flag;
  }
  
  /**
   * Get the value of m_DataSetFirstFirst.
   *
   * @return Value of m_DataSetFirstFirst.
   */
  public boolean getAdvanceDataSetFirst() {
    
    return m_AdvanceDataSetFirst;
  }
  
  /**
   * Set the value of m_AdvanceDataSetFirst.
   *
   * @param newAdvanceDataSetFirst Value to assign to m_AdvanceRunFirst.
   */
  public void setAdvanceDataSetFirst(boolean newAdvanceDataSetFirst) {
    
    m_AdvanceDataSetFirst = newAdvanceDataSetFirst;
  }
  
  /**
   * Gets whether the custom property iterator should be used.
   *
   * @return true if so
   */
  public boolean getUsePropertyIterator() {
    
    return m_UsePropertyIterator;
  }

  /**
   * Sets whether the custom property iterator should be used.
   *
   * @param newUsePropertyIterator true if so
   */
  public void setUsePropertyIterator(boolean newUsePropertyIterator) {
    
    m_UsePropertyIterator = newUsePropertyIterator;
  }

  /**
   * Gets the path of properties taken to get to the custom property
   * to iterate over.
   *
   * @return an array of PropertyNodes
   */
  public PropertyNode [] getPropertyPath() {
    
    return m_PropertyPath;
  }
  
  /**
   * Sets the path of properties taken to get to the custom property
   * to iterate over.
   *
   * @param newPropertyPath an array of PropertyNodes
   */
  public void setPropertyPath(PropertyNode [] newPropertyPath) {
    
    m_PropertyPath = newPropertyPath;
  }
  
  /**
   * Sets the array of values to set the custom property to.
   *
   * @param newPropArray a value of type Object which should be an
   * array of the appropriate values.
   */
  public void setPropertyArray(Object newPropArray) {

    m_PropertyArray = newPropArray;
  }

  /**
   * Gets the array of values to set the custom property to.
   *
   * @return a value of type Object which should be an
   * array of the appropriate values.
   */
  public Object getPropertyArray() {

    return m_PropertyArray;
  }

  /**
   * Gets the number of custom iterator values that have been defined
   * for the experiment.
   *
   * @return the number of custom property iterator values.
   */
  public int getPropertyArrayLength() {

    return Array.getLength(m_PropertyArray);
  }

  /**
   * Gets a specified value from the custom property iterator array.
   *
   * @param index the index of the value wanted
   * @return the property array value
   */
  public Object getPropertyArrayValue(int index) {

    return Array.get(m_PropertyArray, index);
  }
  
  /* These may potentially want to be made un-transient if it is decided
   * that experiments may be saved mid-run and later resumed
   */
  /** The current run number when the experiment is running */
  protected transient int m_RunNumber;
  /** The current dataset number when the experiment is running */
  protected transient int m_DatasetNumber;
  /** The current custom property value index when the experiment is running */
  protected transient int m_PropertyNumber;
  /** True if the experiment has finished running */
  protected transient boolean m_Finished = true;
  /** The dataset currently being used */
  protected transient Instances m_CurrentInstances;
  /** The custom property value that has actually been set */
  protected transient int m_CurrentProperty;

  /**
   * When an experiment is running, this returns the current run number.
   *
   * @return the current run number.
   */
  public int getCurrentRunNumber() {
    return m_RunNumber;
  }

  /**
   * When an experiment is running, this returns the current dataset number.
   *
   * @return the current dataset number.
   */
  public int getCurrentDatasetNumber() {
    return m_DatasetNumber;
  }

  /**
   * When an experiment is running, this returns the index of the
   * current custom property value.
   *
   * @return the index of the current custom property value.
   */
  public int getCurrentPropertyNumber() {
    return m_PropertyNumber;
  }
  
  /**
   * Prepares an experiment for running, initializing current iterator
   * settings.
   *
   * @throws Exception if an error occurs
   */
  public void initialize() throws Exception {
    
    m_RunNumber = getRunLower();
    m_DatasetNumber = 0;
    m_PropertyNumber = 0;
    m_CurrentProperty = -1;
    m_CurrentInstances = null;
    m_Finished = false;
    if (m_UsePropertyIterator && (m_PropertyArray == null)) {
      throw new Exception("Null array for property iterator");
    }
    if (getRunLower() > getRunUpper()) {
      throw new Exception("Lower run number is greater than upper run number");
    }
    if (getDatasets().size() == 0) {
      throw new Exception("No datasets have been specified");
    }
    if (m_ResultProducer == null) {
      throw new Exception("No ResultProducer set");
    }
    if (m_ResultListener == null) {
      throw new Exception("No ResultListener set");
    }

    //    if (m_UsePropertyIterator && (m_PropertyArray != null)) {