wekaclassificationbuildapply.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 Carsten Weisse
  * @author Michael Thess
  * @version 1.0
  */

package com.prudsys.pdm.Examples;

import com.prudsys.pdm.Core.CategoricalAttribute;
import com.prudsys.pdm.Core.Category;
import com.prudsys.pdm.Core.MiningAlgorithm;
import com.prudsys.pdm.Core.MiningAlgorithmSpecification;
import com.prudsys.pdm.Core.MiningAttribute;
import com.prudsys.pdm.Core.MiningDataSpecification;
import com.prudsys.pdm.Core.MiningException;
import com.prudsys.pdm.Core.MiningModel;
import com.prudsys.pdm.Input.MiningArrayStream;
import com.prudsys.pdm.Input.MiningFilterStream;
import com.prudsys.pdm.Input.MiningInputStream;
import com.prudsys.pdm.Input.MiningVector;
import com.prudsys.pdm.Input.Records.Arff.MiningArffStream;
import com.prudsys.pdm.Models.Supervised.SupervisedMiningSettings;
import com.prudsys.pdm.Transform.Special.ReplaceMissingValueStream;
import com.prudsys.pdm.Utils.GeneralUtils;

/**
 * Apply Weka classification methods.
 */
public class WekaClassificationBuildApply extends BasisExample {

  /** Should missing values be replaced before calling WEKA methods? */
  private boolean replaceMissingValues = true;

  /**
   * Empty constructor.
   */
  public WekaClassificationBuildApply() {
  }

  /**
   * Run the example of this class.
   *
   * @throws Exception error while example is running
   */
  public void runExample() throws Exception {

    // -----------------------------------------------------------------------
    //  Build Weka classification model
    // -----------------------------------------------------------------------
    // Open data source, (replace missing values), and get metadata:
  	//<<Frank Xu, 27/01/2005
  	//To test scoring funcationality of J48.
    //MiningInputStream inputData0  = new MiningArffStream( "data/arff/soybeanTrain.arff");
  	MiningInputStream inputData0  = new MiningArffStream( "data/arff/iristrain.arff");
  	//Frank Xu, 27/01/2005>>
    MiningInputStream inputData   = inputData0;
    ReplaceMissingValueStream rep = null;
    if (replaceMissingValues) {
      rep       = new ReplaceMissingValueStream(inputData0);
      inputData = new MiningArrayStream( rep.createReplaceMissingValueStream());
    }
    MiningDataSpecification metaData = inputData.getMetaData();

    // Get target attribute:
    MiningAttribute targetAttribute = (MiningAttribute)metaData.getMiningAttribute( "class" );

    // Create MiningSettings object and assign metadata:
    SupervisedMiningSettings miningSettings = new SupervisedMiningSettings();
    miningSettings.setDataSpecification( metaData );

    // Assign settings:
    miningSettings.setTarget(targetAttribute);
    miningSettings.verifySettings();

    // Get default mining algorithm specification from 'algorithms.xml':
    MiningAlgorithmSpecification miningAlgorithmSpecification =
      MiningAlgorithmSpecification.getMiningAlgorithmSpecification( "J48 (Weka)", null );
//        MiningAlgorithmSpecification.getMiningAlgorithmSpecification( "NeuralNetwork (Weka)" );

    if( miningAlgorithmSpecification == null )
      throw new MiningException( "Can't find weka classification method." );

    // Get class name from algorithms specification:
    String className = miningAlgorithmSpecification.getClassname();
    if( className == null )
      throw new MiningException( "className attribute expected." );

    // Set and display mining parameters:
    GeneralUtils.displayMiningAlgSpecParameters(miningAlgorithmSpecification);

    // Create algorithm object with default values:
    MiningAlgorithm algorithm = GeneralUtils.createMiningAlgorithmInstance(className);

    // Put it all together:
    algorithm.setMiningInputStream( inputData );
    algorithm.setMiningSettings( miningSettings );
    algorithm.setMiningAlgorithmSpecification( miningAlgorithmSpecification );
    algorithm.verify();

    // Build the mining model:
    MiningModel model = algorithm.buildModel();
    System.out.println("calculation time [s]: " + algorithm.getTimeSpentToBuildModel());

    // Write to PMML (experimental):
//      FileWriter writer = new FileWriter("data/pmml/WekaClassify.xml");
//      model.writePmml(writer);

    // Show in browser:
//      if (debug == 2) PmmlUtils.openPmmlBrowser("WekaClassify.xml");

    // Read from PMML (experimental):
//      model = new WekaSupervisedMiningModel();
//      FileReader reader = new FileReader("data/pmml/WekaClassify.xml");
//      model.readPmml(reader);

    // -----------------------------------------------------------------------
    //  Apply Weka classification model
    // -----------------------------------------------------------------------
    // Open new input data, (replace missing values), and get meta data:
  	//To test scoring funcationality of J48.
    //inputData0 = new MiningArffStream( "data/arff/soybeanTest.arff");
    inputData0 = new MiningArffStream( "data/arff/iristest.arff");
  	//Frank Xu, 27/01/2005>>
    inputData  = inputData0;
    if (replaceMissingValues)
      inputData  = new MiningFilterStream(inputData0, rep.getMts());

    MiningDataSpecification inputMetaData  = inputData.getMetaData();
    CategoricalAttribute inputTargetAttribute = (CategoricalAttribute)
      inputMetaData.getMiningAttribute( targetAttribute.getName() );

    // Show relation header:
    System.out.println("Prediction:\n");
    for (int i = 0; i < inputMetaData.getAttributesNumber(); i++) {
      System.out.print(inputMetaData.getMiningAttribute(i).getName() + " ");
    };

    // Show classification results:
    System.out.println();
    int i = 0;
    int wrong = 0;
    inputData.reset();
    while (inputData.next()) {
      // Make prediction:
      MiningVector vector = inputData.read();
      double predicted    = model.applyModelFunction(vector);
      Category predTarCat = ((CategoricalAttribute)targetAttribute).getCategory(predicted);

      // Output and stats:
      double realTarCat   = vector.getValue(inputTargetAttribute);
      Category tarCat     = inputTargetAttribute.getCategory(realTarCat);

      System.out.println(" " + ++i +": " + vector + " -> " + predTarCat);
      if (predTarCat == null || ! predTarCat.equals(tarCat) ) {
        wrong = wrong + 1;
      };
    };
    System.out.println("classification rate = " + (100.0 - ((double) wrong / i)*100.0) );
  }

  /**
   * Example of invoking Weka classification methods from XELOPES.
   *
   * @param args arguments (ignored)
   */
  public static void main(String[] args) {

    try {
      new WekaClassificationBuildApply().runExample();
    }
    catch (Exception ex) {
      ex.printStackTrace();
    }
  }

}