iticollaborativefilteringbuild.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 java.io.FileWriter;
import java.util.Vector;

import com.prudsys.pdm.Core.CategoricalAttribute;
import com.prudsys.pdm.Core.Category;
import com.prudsys.pdm.Core.MiningAlgorithmSpecification;
import com.prudsys.pdm.Core.MiningDataSpecification;
import com.prudsys.pdm.Core.MiningException;
import com.prudsys.pdm.Core.MiningModel;
import com.prudsys.pdm.Input.MiningInputStream;
import com.prudsys.pdm.Input.Records.Csv.MiningCsvStream;
import com.prudsys.pdm.Models.AssociationRules.AssociationRulesAlgorithm;
import com.prudsys.pdm.Models.AssociationRules.AssociationRulesMiningModel;
import com.prudsys.pdm.Models.AssociationRules.AssociationRulesSettings;
import com.prudsys.pdm.Models.AssociationRules.ItemSet;
import com.prudsys.pdm.Models.AssociationRules.RuleSet;
import com.prudsys.pdm.Utils.GeneralUtils;
import com.prudsys.pdm.Utils.PmmlUtils;

/**
 * Builds an Item-to-Item Collaborative Filtering model and
 * writes it to PMML file 'ItemToItemCollaborativeFilteringModel.xml'.
 */
public class ITICollaborativeFilteringBuild extends BasisExample {

  /**
   * Empty constructor.
   */
  public ITICollaborativeFilteringBuild() {
    debug = 0;
  }

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

    // Open data source and get metadata:
    MiningInputStream inputData      = new MiningCsvStream( "data/csv/transact.csv" );
    inputData.open();

    MiningDataSpecification metaData = inputData.getMetaData();

    // Get transactional attributes:
    CategoricalAttribute categoryItemId     = (CategoricalAttribute)metaData.getMiningAttribute( "itemId" );
    CategoricalAttribute categoryTransactId = (CategoricalAttribute)metaData.getMiningAttribute( "transactId" );
    categoryTransactId.setUnstoredCategories(true);

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

    // Assign settings:
    miningSettings.setItemId( categoryItemId );
    miningSettings.setTransactionId( categoryTransactId );
    miningSettings.setMinimumSupport( 0.01 );
    miningSettings.setMinimumConfidence( 0 );
    miningSettings.verifySettings();

    // Generate mining algorithm specification directly:
    MiningAlgorithmSpecification miningAlgorithmSpecification =
      MiningAlgorithmSpecification.getMiningAlgorithmSpecification( "ITICollaborativeFiltering", null);
    if( miningAlgorithmSpecification == null )
      throw new MiningException( "Can't find application ITICollaborativeFiltering." );

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

    // Set and display mining parameters:
    miningAlgorithmSpecification.setMAPValue("minimumCosinus", "0");
    miningAlgorithmSpecification.setMAPValue("minimumSupportCount", "1"); // overrides minSupp
    miningAlgorithmSpecification.setMAPValue("minimumItemSize", "1");
    miningAlgorithmSpecification.setMAPValue("maximumItemSize", "-1");
    miningAlgorithmSpecification.setMAPValue("createLargeItemSets", "true");
    GeneralUtils.displayMiningAlgSpecParameters(miningAlgorithmSpecification);

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

    // Put it all together:
    algorithm.setMiningInputStream( inputData );
    algorithm.setMiningSettings( miningSettings );
    algorithm.setMiningAlgorithmSpecification( miningAlgorithmSpecification );
    // Parameter specific for AssociationRulesAlgorithm but not in MAS:
    algorithm.setExportTransactIds(false);
    algorithm.setExportTransactItemNames( AssociationRulesMiningModel.EXPORT_PMML_NAME_TYPE_XELOPES );
    algorithm.verify();

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

    // Show results:
    showRules( (AssociationRulesMiningModel) model);

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

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

  /**
   * Example of building an association rules model using decomposition.
   *
   * @param args arguments (ignored)
   */
  public static void main(String[] args) {

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

  /**
   * Show association rules and large itemsets.
   *
   * @param ruleModel model of basket analysis
   * @exception MiningException cannot show rules
   */
  public static void showRules(AssociationRulesMiningModel ruleModel)
    throws MiningException {

      // Get rules and large itemsets from model:
      Vector rules = ruleModel.getAssociationRules();
      Vector LITS  = ruleModel.getLargeItemSets();

      // Get item and transaction attributes:
      CategoricalAttribute itemId     = (CategoricalAttribute)( (AssociationRulesSettings) ruleModel.getMiningSettings() ).getItemId();
      // Get number of rules, large itemsets, and transactions:
      int nLITS  = LITS.size();
      int nRules = rules.size();
      int transactsNumber = ruleModel.getNumberOfTransactions();

      // Show all association rules:
      System.out.println();
      System.out.println("Number of association rules found: " + nRules);
      for (int i = 0; i < nRules; i++) {
        // New rule:
        System.out.print(i + ": ");

        // Get and show rule:
        RuleSet rs   = (RuleSet) rules.elementAt(i);
        int itemSize = rs.getSize();

        // Premise part of rule:
        ItemSet is   = rs.getPremise();
        int nprem    = rs.getPremise().getSize();
        for (int j = 0; j < nprem; j++) {
          int pN        = is.getItemAt(j);
          Category cat  = (Category) itemId.getCategory(pN);
          System.out.print(cat.getValue() + " ");
        };
        System.out.print("=> ");

        // Conclusion part of rule:
        for (int j = nprem; j < itemSize; j++) {
          int pN        = rs.getConclusion().getItemAt(j-nprem);
          Category cat  = (Category) itemId.getCategory(pN);
          System.out.print(cat.getValue() + " ");
        }

        // Show support and confidence of rule:
        double Support    = rs.getSupport() * 100.0;
        double Confidence = rs.getConfidence() * 100.0;
        System.out.print("Supp = " + Math.round(Support*100)/100.0 + ", ");
        System.out.print("Conf = " + Math.round(Confidence*100)/100.0 + ", ");

        // Additional measures:
        ruleModel.buildLargeItemSets();
        double Lift     = ruleModel.lift(rs);
        double Cosine   = ruleModel.cosine(rs);
        System.out.print("Lift = " + Math.round(Lift*100)/100.0 + ", ");
        System.out.println("Cos = " + Math.round(Cosine*100)/100.0);
      }

      // Show large itemsets:
      System.out.println();
      System.out.println("Number of large itemsets found: " + nLITS);
      for (int i = 0; i < nLITS; i++) {
        // New large itemset:
        System.out.print(i + ": ");

        // Get and show large itemset:
        ItemSet is   = (ItemSet) LITS.elementAt(i);
        int itemSize = is.getSize();
        for (int j = 0; j < itemSize; j++) {
          int pN       = is.getItemAt(j);
          Category cat  = (Category) itemId.getCategory(pN);
          System.out.print(cat.getValue() + " ");
        };

        // Show support of large itemset:
        double Support = 100.0 * ((double) is.getSupportCount()) /
                                 transactsNumber;
        System.out.println(" Supp = " + Math.round(Support*100)/100.0);
      }
  }
}