principalcomponents.java

来自「Weka」· Java 代码 · 共 803 行 · 第 1/2 页

/*
 *    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.
 */

/*
 * PrincipalComponents.java
 * Copyright (C) 2007 University of Waikato, Hamilton, New Zealand
 */

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.OptionHandler;
import weka.core.SparseInstance;
import weka.core.Utils;
import weka.core.Capabilities.Capability;
import weka.core.matrix.EigenvalueDecomposition;
import weka.core.matrix.Matrix;
import weka.filters.Filter;
import weka.filters.UnsupervisedFilter;

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

/**
 <!-- globalinfo-start -->
 * Performs a principal components analysis and transformation of the data.<br/>
 * Dimensionality reduction is accomplished by choosing enough eigenvectors to account for some percentage of the variance in the original data -- default 0.95 (95%).<br/>
 * Based on code of the attribute selection scheme 'PrincipalComponents' by Mark Hall and Gabi Schmidberger.
 * <p/>
 <!-- globalinfo-end -->
 * 
 <!-- options-start -->
 * Valid options are: <p/>
 * 
 * <pre> -D
 *  Don't normalize input data.</pre>
 * 
 * <pre> -R &lt;num&gt;
 *  Retain enough PC attributes to account
 *  for this proportion of variance in the original data.
 *  (default: 0.95)</pre>
 * 
 * <pre> -A &lt;num&gt;
 *  Maximum number of attributes to include in 
 *  transformed attribute names.
 *  (-1 = include all, default: 5)</pre>
 * 
 * <pre> -M &lt;num&gt;
 *  Maximum number of PC attributes to retain.
 *  (-1 = include all, default: -1)</pre>
 * 
 <!-- options-end -->
 *
 * @author Mark Hall ([e-mail]) -- attribute selection code
 * @author Gabi Schmidberger ([e-mail]) -- attribute selection code
 * @author fracpete (fracpete at waikato dot ac dot nz) -- filter code
 * @version $Revision: 1.3 $
 */
public class PrincipalComponents
  extends Filter
  implements OptionHandler, UnsupervisedFilter {

  /** for serialization. */
  private static final long serialVersionUID = 4626939780964387784L;

  /** The data to transform analyse/transform. */
  protected Instances m_TrainInstances;

  /** Keep a copy for the class attribute (if set). */
  protected Instances m_TrainCopy;

  /** The header for the transformed data format. */
  protected Instances m_TransformedFormat;

  /** Data has a class set. */
  protected boolean m_HasClass;

  /** Class index. */
  protected int m_ClassIndex;

  /** Number of attributes. */
  protected int m_NumAttribs;

  /** Number of instances. */
  protected int m_NumInstances;

  /** Correlation matrix for the original data. */
  protected double[][] m_Correlation;

  /** Will hold the unordered linear transformations of the (normalized)
      original data. */
  protected double[][] m_Eigenvectors;

  /** Eigenvalues for the corresponding eigenvectors. */
  protected double[] m_Eigenvalues = null;

  /** Sorted eigenvalues. */
  protected int[] m_SortedEigens;

  /** sum of the eigenvalues. */
  protected double m_SumOfEigenValues = 0.0;

  /** Filters for replacing missing values. */
  protected ReplaceMissingValues m_ReplaceMissingFilter;
  
  /** Filter for normalizing the data. */
  protected Normalize m_NormalizeFilter;
  
  /** Filter for turning nominal values into numeric ones. */
  protected NominalToBinary m_NominalToBinaryFilter;
  
  /** Filter for removing class attribute, nominal attributes with 0 or 1 value. */
  protected Remove m_AttributeFilter;

  /** The number of attributes in the pc transformed data. */
  protected int m_OutputNumAtts = -1;

  /** normalize the input data? */
  protected boolean m_Normalize = true;

  /** the amount of varaince to cover in the original data when
      retaining the best n PC's. */
  protected double m_CoverVariance = 0.95;

  /** maximum number of attributes in the transformed attribute name. */
  protected int m_MaxAttrsInName = 5;

  /** maximum number of attributes in the transformed data (-1 for all). */
  protected int m_MaxAttributes = -1;

  /**
   * Returns a string describing this filter.
   *
   * @return 		a description of the filter suitable for
   * 			displaying in the explorer/experimenter gui
   */
  public String globalInfo() {
    return 
        "Performs a principal components analysis and transformation of "
      + "the data.\n"
      + "Dimensionality reduction is accomplished by choosing enough eigenvectors "
      + "to account for some percentage of the variance in the original data -- "
      + "default 0.95 (95%).\n"
      + "Based on code of the attribute selection scheme 'PrincipalComponents' "
      + "by Mark Hall and Gabi Schmidberger.";
  }

  /**
   * Returns an enumeration describing the available options.
   *
   * @return 		an enumeration of all the available options.
   */
  public Enumeration listOptions() {
    Vector result = new Vector();

    result.addElement(new Option(
	"\tDon't normalize input data.", 
	"D", 0, "-D"));

    result.addElement(new Option(
	"\tRetain enough PC attributes to account\n"
	+"\tfor this proportion of variance in the original data.\n"
	+ "\t(default: 0.95)",
	"R", 1, "-R <num>"));

    result.addElement(new Option(
	"\tMaximum number of attributes to include in \n"
	+ "\ttransformed attribute names.\n"
	+ "\t(-1 = include all, default: 5)", 
	"A", 1, "-A <num>"));

    result.addElement(new Option(
	"\tMaximum number of PC attributes to retain.\n"
	+ "\t(-1 = include all, default: -1)", 
	"M", 1, "-M <num>"));

    return result.elements();
  }

  /**
   * Parses a list of options for this object. <p/>
   *
   <!-- options-start -->
   * Valid options are: <p/>
   * 
   * <pre> -D
   *  Don't normalize input data.</pre>
   * 
   * <pre> -R &lt;num&gt;
   *  Retain enough PC attributes to account
   *  for this proportion of variance in the original data.
   *  (default: 0.95)</pre>
   * 
   * <pre> -A &lt;num&gt;
   *  Maximum number of attributes to include in 
   *  transformed attribute names.
   *  (-1 = include all, default: 5)</pre>
   * 
   * <pre> -M &lt;num&gt;
   *  Maximum number of PC attributes to retain.
   *  (-1 = include all, default: -1)</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 {
    String        tmpStr;

    tmpStr = Utils.getOption('R', options);
    if (tmpStr.length() != 0)
      setVarianceCovered(Double.parseDouble(tmpStr));
    else
      setVarianceCovered(0.95);

    tmpStr = Utils.getOption('A', options);
    if (tmpStr.length() != 0)
      setMaximumAttributeNames(Integer.parseInt(tmpStr));
    else
      setMaximumAttributeNames(5);

    tmpStr = Utils.getOption('M', options);
    if (tmpStr.length() != 0)
      setMaximumAttributes(Integer.parseInt(tmpStr));
    else
      setMaximumAttributes(-1);

    setNormalize(!Utils.getFlag('D', options));
  }

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

    result = new Vector<String>();

    result.add("-R");
    result.add("" + getVarianceCovered());

    result.add("-A");
    result.add("" + getMaximumAttributeNames());

    result.add("-M");
    result.add("" + getMaximumAttributes());

    if (!getNormalize())
      result.add("-D");

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

  /**
   * Returns the tip text for this property.
   * 
   * @return 		tip text for this property suitable for
   * 			displaying in the explorer/experimenter gui
   */
  public String normalizeTipText() {
    return "Normalize input data.";
  }

  /**
   * Set whether input data will be normalized.
   * 
   * @param value 	true if input data is to be normalized
   */
  public void setNormalize(boolean value) {
    m_Normalize = value;
  }

  /**
   * Gets whether or not input data is to be normalized.
   * 
   * @return 		true if input data is to be normalized
   */
  public boolean getNormalize() {
    return m_Normalize;
  }

  /**
   * Returns the tip text for this property.
   * 
   * @return 		tip text for this property suitable for
   * 			displaying in the explorer/experimenter gui
   */
  public String varianceCoveredTipText() {
    return "Retain enough PC attributes to account for this proportion of variance.";
  }

  /**
   * Sets the amount of variance to account for when retaining
   * principal components.
   * 
   * @param value 	the proportion of total variance to account for
   */
  public void setVarianceCovered(double value) {
    m_CoverVariance = value;
  }

  /**
   * Gets the proportion of total variance to account for when
   * retaining principal components.
   * 
   * @return 		the proportion of variance to account for
   */
  public double getVarianceCovered() {
    return m_CoverVariance;
  }

  /**
   * Returns the tip text for this property.
   * 
   * @return 		tip text for this property suitable for
   * 			displaying in the explorer/experimenter gui
   */
  public String maximumAttributeNamesTipText() {
    return "The maximum number of attributes to include in transformed attribute names.";
  }

  /**
   * Sets maximum number of attributes to include in
   * transformed attribute names.
   * 
   * @param value 	the maximum number of attributes
   */
  public void setMaximumAttributeNames(int value) {
    m_MaxAttrsInName = value;
  }

  /**
   * Gets maximum number of attributes to include in
   * transformed attribute names.
   * 
   * @return 		the maximum number of attributes
   */
  public int getMaximumAttributeNames() {
    return m_MaxAttrsInName;
  }

  /**
   * Returns the tip text for this property.
   * 
   * @return 		tip text for this property suitable for
   * 			displaying in the explorer/experimenter gui
   */
  public String maximumAttributesTipText() {
    return "The maximum number of PC attributes to retain.";
  }

  /**
   * Sets maximum number of PC attributes to retain.
   * 
   * @param value 	the maximum number of attributes
   */
  public void setMaximumAttributes(int value) {
    m_MaxAttributes = value;
  }

  /**
   * Gets maximum number of PC attributes to retain.
   * 
   * @return 		the maximum number of attributes
   */
  public int getMaximumAttributes() {
    return m_MaxAttributes;
  }

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

    // attributes
    result.enable(Capability.NOMINAL_ATTRIBUTES);
    result.enable(Capability.NUMERIC_ATTRIBUTES);