discretize.java

代码是一个分类器的实现,其中使用了部分weka的源代码。可以将项目导入eclipse运行

/*
 *    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.
 */

/*
 *    Discretize.java
 *    Copyright (C) 1999 Eibe Frank,Len Trigg
 *
 */


package weka.filters.unsupervised.attribute;

import weka.core.Attribute;
import weka.core.Capabilities;
import weka.core.FastVector;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.Option;
import weka.core.Range;
import weka.core.SparseInstance;
import weka.core.Utils;
import weka.core.WeightedInstancesHandler;
import weka.core.Capabilities.Capability;
import weka.filters.UnsupervisedFilter;

import java.util.Enumeration;
import java.util.Vector;

/** 
 <!-- globalinfo-start -->
 * An instance filter that discretizes a range of numeric attributes in the dataset into nominal attributes. Discretization is by simple binning. Skips the class attribute if set.
 * <p/>
 <!-- globalinfo-end -->
 * 
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -unset-class-temporarily
 *  Unsets the class index temporarily before the filter is
 *  applied to the data.
 *  (default: no)</pre>
 * 
 * <pre> -B &lt;num&gt;
 *  Specifies the (maximum) number of bins to divide numeric attributes into.
 *  (default = 10)</pre>
 * 
 * <pre> -M &lt;num&gt;
 *  Specifies the desired weight of instances per bin for
 *  equal-frequency binning. If this is set to a positive
 *  number then the -B option will be ignored.
 *  (default = -1)</pre>
 * 
 * <pre> -F
 *  Use equal-frequency instead of equal-width discretization.</pre>
 * 
 * <pre> -O
 *  Optimize number of bins using leave-one-out estimate
 *  of estimated entropy (for equal-width discretization).
 *  If this is set then the -B option will be ignored.</pre>
 * 
 * <pre> -R &lt;col1,col2-col4,...&gt;
 *  Specifies list of columns to Discretize. First and last are valid indexes.
 *  (default: first-last)</pre>
 * 
 * <pre> -V
 *  Invert matching sense of column indexes.</pre>
 * 
 * <pre> -D
 *  Output binary attributes for discretized attributes.</pre>
 * 
 <!-- options-end -->
 *
 * @author Len Trigg ([e-mail])
 * @author Eibe Frank ([e-mail])
 * @version $Revision: 1.11 $
 */
public class Discretize 
  extends PotentialClassIgnorer 
  implements UnsupervisedFilter, WeightedInstancesHandler {
  
  /** for serialization */
  static final long serialVersionUID = -1358531742174527279L;

  /** Stores which columns to Discretize */
  protected Range m_DiscretizeCols = new Range();

  /** The number of bins to divide the attribute into */
  protected int m_NumBins = 10;

  /** The desired weight of instances per bin */
  protected double m_DesiredWeightOfInstancesPerInterval = -1;

  /** Store the current cutpoints */
  protected double [][] m_CutPoints = null;

  /** Output binary attributes for discretized attributes. */
  protected boolean m_MakeBinary = false;

  /** Find the number of bins using cross-validated entropy. */
  protected boolean m_FindNumBins = false;

  /** Use equal-frequency binning if unsupervised discretization turned on */
  protected boolean m_UseEqualFrequency = false;

  /** The default columns to discretize */
  protected String m_DefaultCols;

  /** Constructor - initialises the filter */
  public Discretize() {

    m_DefaultCols = "first-last";
    setAttributeIndices("first-last");
  }

  /** 
   * Another constructor, sets the attribute indices immediately
   * 
   * @param cols the attribute indices
   */
  public Discretize(String cols) {

    m_DefaultCols = cols;
    setAttributeIndices(cols);
  }

  /**
   * Gets an enumeration describing the available options.
   *
   * @return an enumeration of all the available options.
   */
  public Enumeration listOptions() {
    Vector result = new Vector();
    Enumeration enm = super.listOptions();
    while (enm.hasMoreElements())
      result.add(enm.nextElement());
      
    result.addElement(new Option(
	"\tSpecifies the (maximum) number of bins to divide numeric"
	+ " attributes into.\n"
	+ "\t(default = 10)",
	"B", 1, "-B <num>"));
    
    result.addElement(new Option(
	"\tSpecifies the desired weight of instances per bin for\n"
	+ "\tequal-frequency binning. If this is set to a positive\n"
	+ "\tnumber then the -B option will be ignored.\n"
	+ "\t(default = -1)",
	"M", 1, "-M <num>"));
    
    result.addElement(new Option(
	"\tUse equal-frequency instead of equal-width discretization.",
	"F", 0, "-F"));
    
    result.addElement(new Option(
	"\tOptimize number of bins using leave-one-out estimate\n"+
	"\tof estimated entropy (for equal-width discretization).\n"+
	"\tIf this is set then the -B option will be ignored.",
	"O", 0, "-O"));
    
    result.addElement(new Option(
	"\tSpecifies list of columns to Discretize. First"
	+ " and last are valid indexes.\n"
	+ "\t(default: first-last)",
	"R", 1, "-R <col1,col2-col4,...>"));
    
    result.addElement(new Option(
	"\tInvert matching sense of column indexes.",
	"V", 0, "-V"));
    
    result.addElement(new Option(
	"\tOutput binary attributes for discretized attributes.",
	"D", 0, "-D"));

    return result.elements();
  }


  /**
   * Parses a given list of options. <p/>
   * 
   <!-- options-start -->
   * Valid options are: <p/>
   * 
   * <pre> -unset-class-temporarily
   *  Unsets the class index temporarily before the filter is
   *  applied to the data.
   *  (default: no)</pre>
   * 
   * <pre> -B &lt;num&gt;
   *  Specifies the (maximum) number of bins to divide numeric attributes into.
   *  (default = 10)</pre>
   * 
   * <pre> -M &lt;num&gt;
   *  Specifies the desired weight of instances per bin for
   *  equal-frequency binning. If this is set to a positive
   *  number then the -B option will be ignored.
   *  (default = -1)</pre>
   * 
   * <pre> -F
   *  Use equal-frequency instead of equal-width discretization.</pre>
   * 
   * <pre> -O
   *  Optimize number of bins using leave-one-out estimate
   *  of estimated entropy (for equal-width discretization).
   *  If this is set then the -B option will be ignored.</pre>
   * 
   * <pre> -R &lt;col1,col2-col4,...&gt;
   *  Specifies list of columns to Discretize. First and last are valid indexes.
   *  (default: first-last)</pre>
   * 
   * <pre> -V
   *  Invert matching sense of column indexes.</pre>
   * 
   * <pre> -D
   *  Output binary attributes for discretized attributes.</pre>
   * 
   <!-- options-end -->
   *
   * @param options the list of options as an array of strings
   * @throws Exception if an option is not supported
   */
  public void setOptions(String[] options) throws Exception {

    super.setOptions(options);

    setMakeBinary(Utils.getFlag('D', options));
    setUseEqualFrequency(Utils.getFlag('F', options));
    setFindNumBins(Utils.getFlag('O', options));
    setInvertSelection(Utils.getFlag('V', options));

    String weight = Utils.getOption('M', options);
    if (weight.length() != 0) {
      setDesiredWeightOfInstancesPerInterval((new Double(weight)).doubleValue());
    } else {
      setDesiredWeightOfInstancesPerInterval(-1);
    }

    String numBins = Utils.getOption('B', options);
    if (numBins.length() != 0) {
      setBins(Integer.parseInt(numBins));
    } else {
      setBins(10);
    }
    
    String convertList = Utils.getOption('R', options);
    if (convertList.length() != 0) {
      setAttributeIndices(convertList);
    } else {
      setAttributeIndices(m_DefaultCols);
    }

    if (getInputFormat() != null) {
      setInputFormat(getInputFormat());
    }
  }

  /**
   * Gets the current settings of the filter.
   *
   * @return an array of strings suitable for passing to setOptions
   */
  public String [] getOptions() {
    Vector        result;
    String[]      options;
    int           i;

    result = new Vector();

    options = super.getOptions();
    for (i = 0; i < options.length; i++)
      result.add(options[i]);

    if (getMakeBinary())
      result.add("-D");
    
    if (getUseEqualFrequency())
      result.add("-F");
    
    if (getFindNumBins())
      result.add("-O");
    
    if (getInvertSelection())
      result.add("-V");
    
    result.add("-B");
    result.add("" + getBins());
    
    result.add("-M");
    result.add("" + getDesiredWeightOfInstancesPerInterval());
    
    if (!getAttributeIndices().equals("")) {
      result.add("-R");
      result.add(getAttributeIndices());
    }

    return (String[]) result.toArray(new String[result.size()]);
  }

  /** 
   * Returns the Capabilities of this filter.
   *
   * @return            the capabilities of this object
   * @see               Capabilities
   */
  public Capabilities getCapabilities() {
    Capabilities result = super.getCapabilities();

    // attributes
    result.enableAllAttributes();
    result.enable(Capability.MISSING_VALUES);
    
    // class
    result.enableAllClasses();
    result.enable(Capability.MISSING_CLASS_VALUES);
    if (!getMakeBinary())
      result.enable(Capability.NO_CLASS);
    
    return result;
  }

  /**
   * Sets the format of the input instances.
   *
   * @param instanceInfo an Instances object containing the input instance
   * structure (any instances contained in the object are ignored - only the
   * structure is required).
   * @return true if the outputFormat may be collected immediately
   * @throws Exception if the input format can't be set successfully
   */
  public boolean setInputFormat(Instances instanceInfo) throws Exception {

    if (m_MakeBinary && m_IgnoreClass) {
      throw new IllegalArgumentException("Can't ignore class when " +
					 "changing the number of attributes!");
    }

    super.setInputFormat(instanceInfo);

    m_DiscretizeCols.setUpper(instanceInfo.numAttributes() - 1);
    m_CutPoints = null;
    
    if (getFindNumBins() && getUseEqualFrequency()) {
      throw new IllegalArgumentException("Bin number optimization in conjunction "+
					 "with equal-frequency binning not implemented.");
    }

    // If we implement loading cutfiles, then load 
    //them here and set the output format
    return false;
  }

  /**
   * Input an instance for filtering. Ordinarily the instance is processed
   * and made available for output immediately. Some filters require all
   * instances be read before producing output.
   *
   * @param instance the input instance
   * @return true if the filtered instance may now be
   * collected with output().
   * @throws IllegalStateException if no input format has been defined.
   */
  public boolean input(Instance instance) {

    if (getInputFormat() == null) {
      throw new IllegalStateException("No input instance format defined");
    }
    if (m_NewBatch) {
      resetQueue();
      m_NewBatch = false;
    }
    
    if (m_CutPoints != null) {
      convertInstance(instance);
      return true;
    }

    bufferInput(instance);
    return false;
  }

  /**
   * Signifies that this batch of input to the filter is finished. If the 
   * filter requires all instances prior to filtering, output() may now 
   * be called to retrieve the filtered instances.
   *
   * @return true if there are instances pending output
   * @throws IllegalStateException if no input structure has been defined
   */
  public boolean batchFinished() {

    if (getInputFormat() == null) {
      throw new IllegalStateException("No input instance format defined");
    }
    if (m_CutPoints == null) {
      calculateCutPoints();

      setOutputFormat();

      // If we implement saving cutfiles, save the cuts here

      // Convert pending input instances
      for(int i = 0; i < getInputFormat().numInstances(); i++) {
	convertInstance(getInputFormat().instance(i));
      }
    } 
    flushInput();

    m_NewBatch = true;
    return (numPendingOutput() != 0);
  }

  /**
   * Returns a string describing this filter
   *
   * @return a description of the filter suitable for
   * displaying in the explorer/experimenter gui
   */
  public String globalInfo() {

    return "An instance filter that discretizes a range of numeric"
      + " attributes in the dataset into nominal attributes."
      + " Discretization is by simple binning. Skips the class"
      + " attribute if set.";
  }
  
  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String findNumBinsTipText() {

    return "Optimize number of equal-width bins using leave-one-out. Doesn't " +
      "work for equal-frequency binning";
  }

  /**
   * Get the value of FindNumBins.
   *
   * @return Value of FindNumBins.
   */
  public boolean getFindNumBins() {
    
    return m_FindNumBins;
  }
  
  /**
   * Set the value of FindNumBins.
   *
   * @param newFindNumBins Value to assign to FindNumBins.
   */
  public void setFindNumBins(boolean newFindNumBins) {
    
    m_FindNumBins = newFindNumBins;
  }
  
  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String makeBinaryTipText() {

    return "Make resulting attributes binary.";
  }

  /**
   * Gets whether binary attributes should be made for discretized ones.
   *
   * @return true if attributes will be binarized
   */
  public boolean getMakeBinary() {

    return m_MakeBinary;
  }

  /** 
   * Sets whether binary attributes should be made for discretized ones.
   *
   * @param makeBinary if binary attributes are to be made
   */
  public void setMakeBinary(boolean makeBinary) {

    m_MakeBinary = makeBinary;
  }
  
  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String desiredWeightOfInstancesPerIntervalTipText() {

    return "Sets the desired weight of instances per interval for " +
      "equal-frequency binning.";
  }
  
  /**
   * Get the DesiredWeightOfInstancesPerInterval value.
   * @return the DesiredWeightOfInstancesPerInterval value.
   */
  public double getDesiredWeightOfInstancesPerInterval() {

    return m_DesiredWeightOfInstancesPerInterval;
  }

  /**
   * Set the DesiredWeightOfInstancesPerInterval value.
   * @param newDesiredNumber The new DesiredNumber value.
   */
  public void setDesiredWeightOfInstancesPerInterval(double newDesiredNumber) {
    
    m_DesiredWeightOfInstancesPerInterval = newDesiredNumber;
  }
  
  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for