Package weka.filters.unsupervised.attribute

Source Code of weka.filters.unsupervised.attribute.NominalToBinary

/*
*    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.
*/

/*
*    NominalToBinary.java
*    Copyright (C) 1999 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.Range;
import weka.core.RevisionUtils;
import weka.core.SparseInstance;
import weka.core.Utils;
import weka.core.Capabilities.Capability;
import weka.filters.Filter;
import weka.filters.UnsupervisedFilter;

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

/**
<!-- globalinfo-start -->
* Converts all nominal attributes into binary numeric attributes. An attribute with k values is transformed into k binary attributes if the class is nominal (using the one-attribute-per-value approach). Binary attributes are left binary, if option '-A' is not given.If the class is numeric, you might want to use the supervised version of this filter.
* <p/>
<!-- globalinfo-end -->
*
<!-- options-start -->
* Valid options are: <p/>
*
* <pre> -N
*  Sets if binary attributes are to be coded as nominal ones.</pre>
*
* <pre> -A
*  For each nominal value a new attribute is created,
*  not only if there are more than 2 values.</pre>
*
* <pre> -R &lt;col1,col2-col4,...&gt;
*  Specifies list of columns to act on. First and last are
*  valid indexes.
*  (default: first-last)</pre>
*
* <pre> -V
*  Invert matching sense of column indexes.</pre>
*
<!-- options-end -->
*
* @author Eibe Frank (eibe@cs.waikato.ac.nz)
* @version $Revision: 1.12 $
*/
public class NominalToBinary
  extends Filter
  implements UnsupervisedFilter, OptionHandler {
 
  /** for serialization */
  static final long serialVersionUID = -1130642825710549138L;

  /** Stores which columns to act on */
  protected Range m_Columns = new Range();

  /** Are the new attributes going to be nominal or numeric ones? */
  private boolean m_Numeric = true;

  /** Are all values transformed into new attributes? */
  private boolean m_TransformAll = false;

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

    setAttributeIndices("first-last");
  }

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

    return "Converts all nominal attributes into binary numeric attributes. An "
      + "attribute with k values is transformed into k binary attributes if "
      + "the class is nominal (using the one-attribute-per-value approach). "
      + "Binary attributes are left binary, if option '-A' is not given."
      + "If the class is numeric, you might want to use the supervised version of "
      + "this filter.";
  }

  /**
   * 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);
    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 {

    super.setInputFormat(instanceInfo);

    m_Columns.setUpper(instanceInfo.numAttributes() - 1);

    setOutputFormat();
    return true;
  }

  /**
   * Input an instance for filtering. Filter requires all
   * training 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 set
   */
  public boolean input(Instance instance) {

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

    convertInstance(instance);
    return true;
  }

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

    Vector newVector = new Vector(3);

    newVector.addElement(new Option(
  "\tSets if binary attributes are to be coded as nominal ones.",
  "N", 0, "-N"));

    newVector.addElement(new Option(
  "\tFor each nominal value a new attribute is created, \n"
  + "\tnot only if there are more than 2 values.",
  "A", 0, "-A"));

    newVector.addElement(new Option(
  "\tSpecifies list of columns to act on. First and last are \n"
  + "\tvalid indexes.\n"
  + "\t(default: first-last)",
  "R", 1, "-R <col1,col2-col4,...>"));

    newVector.addElement(new Option(
  "\tInvert matching sense of column indexes.",
  "V", 0, "-V"));

    return newVector.elements();
  }


  /**
   * Parses a given list of options. <p/>
   *
   <!-- options-start -->
   * Valid options are: <p/>
   *
   * <pre> -N
   *  Sets if binary attributes are to be coded as nominal ones.</pre>
   *
   * <pre> -A
   *  For each nominal value a new attribute is created,
   *  not only if there are more than 2 values.</pre>
   *
   * <pre> -R &lt;col1,col2-col4,...&gt;
   *  Specifies list of columns to act on. First and last are
   *  valid indexes.
   *  (default: first-last)</pre>
   *
   * <pre> -V
   *  Invert matching sense of column indexes.</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 {

    setBinaryAttributesNominal(Utils.getFlag('N', options));

    setTransformAllValues(Utils.getFlag('A', options));

    String convertList = Utils.getOption('R', options);
    if (convertList.length() != 0) {
      setAttributeIndices(convertList);
    } else {
      setAttributeIndices("first-last");
    }
    setInvertSelection(Utils.getFlag('V', options));

    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() {

    String [] options = new String [4];
    int current = 0;

    if (getBinaryAttributesNominal()) {
      options[current++] = "-N";
    }

    if (getTransformAllValues()) {
      options[current++] = "-A";
    }

    if (!getAttributeIndices().equals("")) {
      options[current++] = "-R"; options[current++] = getAttributeIndices();
    }
    if (getInvertSelection()) {
      options[current++] = "-V";
    }

    while (current < options.length) {
      options[current++] = "";
    }
    return options;
  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String binaryAttributesNominalTipText() {
    return "Whether resulting binary attributes will be nominal.";
  }

  /**
   * Gets if binary attributes are to be treated as nominal ones.
   *
   * @return true if binary attributes are to be treated as nominal ones
   */
  public boolean getBinaryAttributesNominal() {

    return !m_Numeric;
  }

  /**
   * Sets if binary attributes are to be treates as nominal ones.
   *
   * @param bool true if binary attributes are to be treated as nominal ones
   */
  public void setBinaryAttributesNominal(boolean bool) {

    m_Numeric = !bool;
  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String transformAllValuesTipText() {
    return "Whether all nominal values are turned into new attributes, not only if there are more than 2.";
  }

  /**
   * Gets if all nominal values are turned into new attributes, not only if
   * there are more than 2.
   *
   * @return true all nominal values are transformed into new attributes
   */
  public boolean getTransformAllValues() {

    return m_TransformAll;
  }

  /**
   * Sets whether all nominal values are transformed into new attributes, not
   * just if there are more than 2.
   *
   * @param bool true if all nominal value are transformed into new attributes
   */
  public void setTransformAllValues(boolean bool) {

    m_TransformAll = bool;
  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String invertSelectionTipText() {

    return "Set attribute selection mode. If false, only selected"
      + " (numeric) attributes in the range will be discretized; if"
      + " true, only non-selected attributes will be discretized.";
  }

  /**
   * Gets whether the supplied columns are to be removed or kept
   *
   * @return true if the supplied columns will be kept
   */
  public boolean getInvertSelection() {

    return m_Columns.getInvert();
  }

  /**
   * Sets whether selected columns should be removed or kept. If true the
   * selected columns are kept and unselected columns are deleted. If false
   * selected columns are deleted and unselected columns are kept.
   *
   * @param invert the new invert setting
   */
  public void setInvertSelection(boolean invert) {

    m_Columns.setInvert(invert);
  }

  /**
   * Returns the tip text for this property
   *
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String attributeIndicesTipText() {
    return "Specify range of attributes to act on."
      + " This is a comma separated list of attribute indices, with"
      + " \"first\" and \"last\" valid values. Specify an inclusive"
      + " range with \"-\". E.g: \"first-3,5,6-10,last\".";
  }

  /**
   * Gets the current range selection
   *
   * @return a string containing a comma separated list of ranges
   */
  public String getAttributeIndices() {

    return m_Columns.getRanges();
  }

  /**
   * Sets which attributes are to be acted on.
   *
   * @param rangeList a string representing the list of attributes. Since
   * the string will typically come from a user, attributes are indexed from
   * 1. <br>
   * eg: first-3,5,6-last
   * @throws IllegalArgumentException if an invalid range list is supplied
   */
  public void setAttributeIndices(String rangeList) {

    m_Columns.setRanges(rangeList);
  }

  /**
   * Set the output format if the class is nominal.
   */
  private void setOutputFormat() {

    FastVector newAtts;
    int newClassIndex;
    StringBuffer attributeName;
    Instances outputFormat;
    FastVector vals;

    // Compute new attributes

    newClassIndex = getInputFormat().classIndex();
    newAtts = new FastVector();
    for (int j = 0; j < getInputFormat().numAttributes(); j++) {
      Attribute att = getInputFormat().attribute(j);
      if (!att.isNominal() || (j == getInputFormat().classIndex()) ||
    !m_Columns.isInRange(j)) {
  newAtts.addElement(att.copy());
      } else {
  if ( (att.numValues() <= 2) && (!m_TransformAll) ) {
    if (m_Numeric) {
      newAtts.addElement(new Attribute(att.name()));
    } else {
      newAtts.addElement(att.copy());
    }
  } else {

    if (newClassIndex >= 0 && j < getInputFormat().classIndex()) {
      newClassIndex += att.numValues() - 1;
    }

    // Compute values for new attributes
    for (int k = 0; k < att.numValues(); k++) {
      attributeName =
        new StringBuffer(att.name() + "=");
      attributeName.append(att.value(k));
      if (m_Numeric) {
        newAtts.
    addElement(new Attribute(attributeName.toString()));
      } else {
        vals = new FastVector(2);
        vals.addElement("f"); vals.addElement("t");
        newAtts.
    addElement(new Attribute(attributeName.toString(), vals));
      }
    }
  }
      }
    }
    outputFormat = new Instances(getInputFormat().relationName(),
         newAtts, 0);
    outputFormat.setClassIndex(newClassIndex);
    setOutputFormat(outputFormat);
  }

  /**
   * Convert a single instance over if the class is nominal. The converted
   * instance is added to the end of the output queue.
   *
   * @param instance the instance to convert
   */
  private void convertInstance(Instance instance) {

    double [] vals = new double [outputFormatPeek().numAttributes()];
    int attSoFar = 0;

    for(int j = 0; j < getInputFormat().numAttributes(); j++) {
      Attribute att = getInputFormat().attribute(j);
      if (!att.isNominal() || (j == getInputFormat().classIndex()) ||
    !m_Columns.isInRange(j)) {
  vals[attSoFar] = instance.value(j);
  attSoFar++;
      } else {
  if ( (att.numValues() <= 2) && (!m_TransformAll) ) {
    vals[attSoFar] = instance.value(j);
    attSoFar++;
  } else {
    if (instance.isMissing(j)) {
      for (int k = 0; k < att.numValues(); k++) {
              vals[attSoFar + k] = instance.value(j);
      }
    } else {
      for (int k = 0; k < att.numValues(); k++) {
        if (k == (int)instance.value(j)) {
                vals[attSoFar + k] = 1;
        } else {
                vals[attSoFar + k] = 0;
        }
      }
    }
    attSoFar += att.numValues();
  }
      }
    }
    Instance inst = null;
    if (instance instanceof SparseInstance) {
      inst = new SparseInstance(instance.weight(), vals);
    } else {
      inst = new Instance(instance.weight(), vals);
    }
    inst.setDataset(getOutputFormat());
    copyValues(inst, false, instance.dataset(), getOutputFormat());
    inst.setDataset(getOutputFormat());
    push(inst);
  }
 
  /**
   * Returns the revision string.
   *
   * @return    the revision
   */
  public String getRevision() {
    return RevisionUtils.extract("$Revision: 1.12 $");
  }

  /**
   * Main method for testing this class.
   *
   * @param argv should contain arguments to the filter:
   * use -h for help
   */
  public static void main(String [] argv) {
    runFilter(new NominalToBinary(), argv);
  }
}
TOP

Related Classes of weka.filters.unsupervised.attribute.NominalToBinary

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.