categorization.java

一个数据挖掘软件ALPHAMINERR的整个过程的JAVA版源代码

/*
 *    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.
 */

 /**
  * Title: XELOPES Data Mining Library
  * Description: The XELOPES library is an open platform-independent and data-source-independent library for Embedded Data Mining.
  * Copyright: Copyright (c) 2002 Prudential Systems Software GmbH
  * Company: ZSoft (www.zsoft.ru), Prudsys (www.prudsys.com)
  * @author Michael Thess
  * @version 1.0
  */
package com.prudsys.pdm.Transform.OneToOne;

import com.prudsys.pdm.Adapters.PmmlVersion20.DerivedField;
import com.prudsys.pdm.Core.CategoricalAttribute;
import com.prudsys.pdm.Core.Category;
import com.prudsys.pdm.Core.MiningAttribute;
import com.prudsys.pdm.Core.MiningException;
import com.prudsys.pdm.Core.NumericAttribute;
import com.prudsys.pdm.Transform.OneToOneMapping;

import eti.bi.util.NumberFormatter;

/**
 * Realization of categorization.
 *
 * Transforms a numeric into a categorical attribute
 * in the most simple way: The values are used
 * as categories of the new categorical attribute.
 * For this purpose, the double values of the numeric
 * attributes are wrapped as Double which are used
 * as category objects.
 *
 * The categorical attribute is of the type
 * 'unboundedCategories' and the categories are
 * dynamically added in the transformAttributeValue
 * method.
 */
public class Categorization extends OneToOneMapping
{
  // -----------------------------------------------------------------------
  //  Variables declarations
  // -----------------------------------------------------------------------
  /** Transformed attribute. */
  private CategoricalAttribute transformedAttribute;

  // -----------------------------------------------------------------------
  //  Constructor
  // -----------------------------------------------------------------------
  /**
   * Empty constructor.
   */
  public Categorization()
  {
  }

  // -----------------------------------------------------------------------
  //  Transformation methods
  // -----------------------------------------------------------------------
  /**
   * Transforms the numeric source attribute. The result is the
   * categorical target attribute. So far, no categories are
   * added because the values of the numeric attribute are
   * not apriori known. They are dynamically added in the
   * transformAttributeValue method.
   *
   * @return transformed attribute
   * @exception MiningException could not transform attribute
   */
  public MiningAttribute transformAttribute() throws MiningException
  {
      if (getSourceAttribute() == null)
        throw new MiningException("Could not find source attribute: " + sourceName);

      if (! (getSourceAttribute() instanceof NumericAttribute))
        throw new MiningException("Source attribute '" + sourceName + "' must be numeric");

      NumericAttribute sourceAttribute = (NumericAttribute) getSourceAttribute();
      transformedAttribute             = new CategoricalAttribute();

      transformedAttribute.setName( getTargetNameDynamic() );
      transformedAttribute.setUnboundedCategories( true );
      
      return transformedAttribute;
  }

  /**
   * Transforms value of numeric attribute. The result is the key of
   * the categorical attribute. If the value is new, it is
   * the key of the newly created category.
   *
   * @param attributeValue value of attribute to be transformed
   * @return transformed (categorized) value, key of corresponding category
   * @exception MiningException could not transform attribute value
   */
  public double transformAttributeValue( double attributeValue ) throws MiningException
  {
    String formattedString = NumberFormatter.format(attributeValue);
    Category cat  = new Category(formattedString);
    double key    = transformedAttribute.getKey( cat );
    //<<tyleung 23/3/2005
    if ( Category.isMissingValue(key) && !Double.isNaN(attributeValue))
        key = transformedAttribute.addCategory( cat );
    //tyleung 23/3/2005 >>
    return key;
  }

  // -----------------------------------------------------------------------
  //  Methods of PMML handling
  // -----------------------------------------------------------------------
  /**
   * Creates PMML object DerivedField of this object of Categorization type.
   *
   * @return DerivedField element
   * @exception MiningException cannot create PMML object
   * @see com.prudsys.pdm.Adapters.PmmlVersion20.Categorization
   */
  public Object createPmmlObject() throws MiningException
  {
    DerivedField field = (DerivedField) super.createPmmlObject();

    com.prudsys.pdm.Adapters.PmmlVersion20.Categorization categ = new com.prudsys.pdm.Adapters.PmmlVersion20.Categorization();
    categ.setField( sourceName );

    field.setCategorization( categ );
    return field;
  }

  /**
   * Creates this object from PMML object categorization.
   *
   * @param pmml pmml element
   * @exception MiningException cannot parse PMML object
   * @see com.prudsys.pdm.Adapters.PmmlVersion20.Categorization
   */
  public void parsePmmlObject(Object pmml) throws MiningException
  {
    super.parsePmmlObject(pmml);

    DerivedField field = (DerivedField) pmml;

    com.prudsys.pdm.Adapters.PmmlVersion20.Categorization categ = field.getCategorization();
    sourceName = categ.getField();
  }
}