plsfilter.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.
 */

/*
 * PLSFilter.java
 * Copyright (C) 2006 University of Waikato, Hamilton, New Zealand
 *
 */

package weka.filters.supervised.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.SelectedTag;
import weka.core.Tag;
import weka.core.TechnicalInformation;
import weka.core.TechnicalInformationHandler;
import weka.core.Utils;
import weka.core.Capabilities.Capability;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;
import weka.core.matrix.EigenvalueDecomposition;
import weka.core.matrix.Matrix;
import weka.filters.Filter;
import weka.filters.SimpleBatchFilter;
import weka.filters.SupervisedFilter;
import weka.filters.unsupervised.attribute.Center;
import weka.filters.unsupervised.attribute.ReplaceMissingValues;
import weka.filters.unsupervised.attribute.Standardize;

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

/** 
 <!-- globalinfo-start -->
 * Runs Partial Least Square Regression over the given instances and computes the resulting beta matrix for prediction.<br/>
 * By default it replaces missing values and centers the data.<br/>
 * <br/>
 * For more information see:<br/>
 * <br/>
 * Tormod Naes, Tomas Isaksson, Tom Fearn, Tony Davies (2002). A User Friendly Guide to Multivariate Calibration and Classification. NIR Publications.<br/>
 * <br/>
 * StatSoft, Inc.. Partial Least Squares (PLS).<br/>
 * <br/>
 * Bent Jorgensen, Yuri Goegebeur. Module 7: Partial least squares regression I.<br/>
 * <br/>
 * S. de Jong (1993). SIMPLS: an alternative approach to partial least squares regression. Chemometrics and Intelligent Laboratory Systems. 18:251-263.
 * <p/>
 <!-- globalinfo-end -->
 *
 <!-- technical-bibtex-start -->
 * BibTeX:
 * <pre>
 * &#64;book{Naes2002,
 *    author = {Tormod Naes and Tomas Isaksson and Tom Fearn and Tony Davies},
 *    publisher = {NIR Publications},
 *    title = {A User Friendly Guide to Multivariate Calibration and Classification},
 *    year = {2002},
 *    ISBN = {0-9528666-2-5}
 * }
 * 
 * &#64;misc{missing_id,
 *    author = {StatSoft, Inc.},
 *    booktitle = {Electronic Textbook StatSoft},
 *    title = {Partial Least Squares (PLS)},
 *    HTTP = {http://www.statsoft.com/textbook/stpls.html}
 * }
 * 
 * &#64;misc{missing_id,
 *    author = {Bent Jorgensen and Yuri Goegebeur},
 *    booktitle = {ST02: Multivariate Data Analysis and Chemometrics},
 *    title = {Module 7: Partial least squares regression I},
 *    HTTP = {http://statmaster.sdu.dk/courses/ST02/module07/}
 * }
 * 
 * &#64;article{Jong1993,
 *    author = {S. de Jong},
 *    journal = {Chemometrics and Intelligent Laboratory Systems},
 *    pages = {251-263},
 *    title = {SIMPLS: an alternative approach to partial least squares regression},
 *    volume = {18},
 *    year = {1993}
 * }
 * </pre>
 * <p/>
 <!-- technical-bibtex-end -->
 *
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -D
 *  Turns on output of debugging information.</pre>
 * 
 * <pre> -C &lt;num&gt;
 *  The number of components to compute.
 *  (default: 20)</pre>
 * 
 * <pre> -U
 *  Updates the class attribute as well.
 *  (default: off)</pre>
 * 
 * <pre> -M
 *  Turns replacing of missing values on.
 *  (default: off)</pre>
 * 
 * <pre> -A &lt;SIMPLS|PLS1&gt;
 *  The algorithm to use.
 *  (default: PLS1)</pre>
 * 
 * <pre> -P &lt;none|center|standardize&gt;
 *  The type of preprocessing that is applied to the data.
 *  (default: center)</pre>
 * 
 <!-- options-end -->
 *
 * @author FracPete (fracpete at waikato dot ac dot nz)
 * @version $Revision: 1.3 $
 */
public class PLSFilter
  extends SimpleBatchFilter 
  implements SupervisedFilter, TechnicalInformationHandler {

  /** for serialization */
  static final long serialVersionUID = -3335106965521265631L;

  /** the type of algorithm: SIMPLS */
  public static final int ALGORITHM_SIMPLS = 1;
  /** the type of algorithm: PLS1 */
  public static final int ALGORITHM_PLS1 = 2;
  /** the types of algorithm */
  public static final Tag[] TAGS_ALGORITHM = {
    new Tag(ALGORITHM_SIMPLS, "SIMPLS"),
    new Tag(ALGORITHM_PLS1, "PLS1")
  };

  /** the type of preprocessing: None */
  public static final int PREPROCESSING_NONE = 0;
  /** the type of preprocessing: Center */
  public static final int PREPROCESSING_CENTER = 1;
  /** the type of preprocessing: Standardize */
  public static final int PREPROCESSING_STANDARDIZE = 2;
  /** the types of preprocessing */
  public static final Tag[] TAGS_PREPROCESSING = {
    new Tag(PREPROCESSING_NONE, "none"),
    new Tag(PREPROCESSING_CENTER, "center"),
    new Tag(PREPROCESSING_STANDARDIZE, "standardize")
  };

  /** the maximum number of components to generate */
  protected int m_NumComponents = 20;
  
  /** the type of algorithm */
  protected int m_Algorithm = ALGORITHM_PLS1;

  /** the regression vector "r-hat" for PLS1 */
  protected Matrix m_PLS1_RegVector = null;

  /** the P matrix for PLS1 */
  protected Matrix m_PLS1_P = null;

  /** the W matrix for PLS1 */
  protected Matrix m_PLS1_W = null;

  /** the b-hat vector for PLS1 */
  protected Matrix m_PLS1_b_hat = null;
  
  /** the W matrix for SIMPLS */
  protected Matrix m_SIMPLS_W = null;
  
  /** the B matrix for SIMPLS (used for prediction) */
  protected Matrix m_SIMPLS_B = null;
  
  /** whether to include the prediction, i.e., modifying the class attribute */
  protected boolean m_PerformPrediction = false;

  /** for replacing missing values */
  protected Filter m_Missing = null;
  
  /** whether to replace missing values */
  protected boolean m_ReplaceMissing = true;
  
  /** for centering the data */
  protected Filter m_Filter = null;
  
  /** the type of preprocessing */
  protected int m_Preprocessing = PREPROCESSING_CENTER;

  /** the mean of the class */
  protected double m_ClassMean = 0;

  /** the standard deviation of the class */
  protected double m_ClassStdDev = 0;
  
  /**
   * default constructor
   */
  public PLSFilter() {
    super();
    
    // setup pre-processing
    m_Missing = new ReplaceMissingValues();
    m_Filter  = new Center();
  }
  
  /**
   * Returns a string describing this classifier.
   *
   * @return      a description of the classifier suitable for
   *              displaying in the explorer/experimenter gui
   */
  public String globalInfo() {
    return 
        "Runs Partial Least Square Regression over the given instances "
      + "and computes the resulting beta matrix for prediction.\n"
      + "By default it replaces missing values and centers the data.\n\n"
      + "For more information see:\n\n"
      + getTechnicalInformation().toString();
  }

  /**
   * Returns an instance of a TechnicalInformation object, containing 
   * detailed information about the technical background of this class,
   * e.g., paper reference or book this class is based on.
   * 
   * @return the technical information about this class
   */
  public TechnicalInformation getTechnicalInformation() {
    TechnicalInformation 	result;
    TechnicalInformation 	additional;
    
    result = new TechnicalInformation(Type.BOOK);
    result.setValue(Field.AUTHOR, "Tormod Naes and Tomas Isaksson and Tom Fearn and Tony Davies");
    result.setValue(Field.YEAR, "2002");
    result.setValue(Field.TITLE, "A User Friendly Guide to Multivariate Calibration and Classification");
    result.setValue(Field.PUBLISHER, "NIR Publications");
    result.setValue(Field.ISBN, "0-9528666-2-5");
    
    additional = result.add(Type.MISC);
    additional.setValue(Field.AUTHOR, "StatSoft, Inc.");
    additional.setValue(Field.TITLE, "Partial Least Squares (PLS)");
    additional.setValue(Field.BOOKTITLE, "Electronic Textbook StatSoft");
    additional.setValue(Field.HTTP, "http://www.statsoft.com/textbook/stpls.html");
    
    additional = result.add(Type.MISC);
    additional.setValue(Field.AUTHOR, "Bent Jorgensen and Yuri Goegebeur");
    additional.setValue(Field.TITLE, "Module 7: Partial least squares regression I");
    additional.setValue(Field.BOOKTITLE, "ST02: Multivariate Data Analysis and Chemometrics");
    additional.setValue(Field.HTTP, "http://statmaster.sdu.dk/courses/ST02/module07/");
    
    additional = result.add(Type.ARTICLE);
    additional.setValue(Field.AUTHOR, "S. de Jong");
    additional.setValue(Field.YEAR, "1993");
    additional.setValue(Field.TITLE, "SIMPLS: an alternative approach to partial least squares regression");
    additional.setValue(Field.JOURNAL, "Chemometrics and Intelligent Laboratory Systems");
    additional.setValue(Field.VOLUME, "18");
    additional.setValue(Field.PAGES, "251-263");
    
    return result;
  }

  /**
   * Gets an enumeration describing the available options.
   *
   * @return an enumeration of all the available options.
   */
  public Enumeration listOptions() {
    Vector		result;
    Enumeration		enm;
    String		param;
    SelectedTag		tag;
    int			i;

    result = new Vector();

    enm = super.listOptions();
    while (enm.hasMoreElements())
      result.addElement(enm.nextElement());

    result.addElement(new Option(
	"\tThe number of components to compute.\n"
	+ "\t(default: 20)",
	"C", 1, "-C <num>"));

    result.addElement(new Option(
	"\tUpdates the class attribute as well.\n"
	+ "\t(default: off)",
	"U", 0, "-U"));

    result.addElement(new Option(
	"\tTurns replacing of missing values on.\n"
	+ "\t(default: off)",
	"M", 0, "-M"));

    param = "";
    for (i = 0; i < TAGS_ALGORITHM.length; i++) {
      if (i > 0)
	param += "|";
      tag = new SelectedTag(TAGS_ALGORITHM[i].getID(), TAGS_ALGORITHM);
      param += tag.getSelectedTag().getReadable();
    }
    result.addElement(new Option(
	"\tThe algorithm to use.\n"
	+ "\t(default: PLS1)",
	"A", 1, "-A <" + param + ">"));

    param = "";
    for (i = 0; i < TAGS_PREPROCESSING.length; i++) {
      if (i > 0)
	param += "|";
      tag = new SelectedTag(TAGS_PREPROCESSING[i].getID(), TAGS_PREPROCESSING);
      param += tag.getSelectedTag().getReadable();
    }
    result.addElement(new Option(
	"\tThe type of preprocessing that is applied to the data.\n"
	+ "\t(default: center)",
	"P", 1, "-P <" + param + ">"));

    return result.elements();
  }

  /**
   * returns the options of the current setup
   *
   * @return      the current options
   */
  public String[] getOptions() {
    int       i;
    Vector    result;
    String[]  options;

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

    result.add("-C");
    result.add("" + getNumComponents());

    if (getPerformPrediction())
      result.add("-U");
    
    if (getReplaceMissing())
      result.add("-M");
    
    result.add("-A");
    result.add("" + getAlgorithm().getSelectedTag().getReadable());

    result.add("-P");
    result.add("" + getPreprocessing().getSelectedTag().getReadable());

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

  /**
   * Parses the options for this object. <p/>
   *
   <!-- options-start -->
   * Valid options are: <p/>
   * 
   * <pre> -D
   *  Turns on output of debugging information.</pre>
   * 
   * <pre> -C &lt;num&gt;
   *  The number of components to compute.
   *  (default: 20)</pre>
   * 
   * <pre> -U