checkattributeselection.java

这是关于数据挖掘的一些算法

   * @return index 0 is true if the test was passed, index 1 is true if test 
   *         was acceptable
   * @see TestInstances#CLASS_IS_LAST
   */
  protected boolean[] canHandleClassAsNthAttribute(
      boolean nominalPredictor,
      boolean numericPredictor, 
      boolean stringPredictor, 
      boolean datePredictor,
      boolean relationalPredictor,
      boolean multiInstance,
      int classType,
      int classIndex) {
    
    if (classIndex == TestInstances.CLASS_IS_LAST)
      print("class attribute as last attribute");
    else
      print("class attribute as " + (classIndex + 1) + ". attribute");
    printAttributeSummary(
        nominalPredictor, numericPredictor, stringPredictor, datePredictor, relationalPredictor, multiInstance, classType);
    print("...");
    FastVector accepts = new FastVector();
    int numTrain = getNumInstances(), numClasses = 2, missingLevel = 0;
    boolean predictorMissing = false, classMissing = false;
    
    return runBasicTest(nominalPredictor, numericPredictor, stringPredictor, 
                        datePredictor, relationalPredictor, 
                        multiInstance,
                        classType,
                        classIndex,
                        missingLevel, predictorMissing, classMissing,
                        numTrain, numClasses, 
                        accepts);
  }
  
  /**
   * Checks whether the scheme can handle zero training instances.
   *
   * @param nominalPredictor if true use nominal predictor attributes
   * @param numericPredictor if true use numeric predictor attributes
   * @param stringPredictor if true use string predictor attributes
   * @param datePredictor if true use date predictor attributes
   * @param relationalPredictor if true use relational predictor attributes
   * @param multiInstance whether multi-instance is needed
   * @param classType the class type (NUMERIC, NOMINAL, etc.)
   * @return index 0 is true if the test was passed, index 1 is true if test 
   *         was acceptable
   */
  protected boolean[] canHandleZeroTraining(
      boolean nominalPredictor,
      boolean numericPredictor, 
      boolean stringPredictor, 
      boolean datePredictor,
      boolean relationalPredictor,
      boolean multiInstance,
      int classType) {
    
    print("handle zero training instances");
    printAttributeSummary(
        nominalPredictor, numericPredictor, stringPredictor, datePredictor, relationalPredictor, multiInstance, classType);
    print("...");
    FastVector accepts = new FastVector();
    accepts.addElement("train");
    accepts.addElement("value");
    int numTrain = 0, numClasses = 2, missingLevel = 0;
    boolean predictorMissing = false, classMissing = false;
    
    return runBasicTest(
              nominalPredictor, numericPredictor, stringPredictor, 
              datePredictor, relationalPredictor, 
              multiInstance,
              classType, 
              missingLevel, predictorMissing, classMissing,
              numTrain, numClasses, 
              accepts);
  }
  
  /**
   * Checks whether the scheme correctly initialises models when 
   * ASSearch.search is called. This test calls search with
   * one training dataset. ASSearch is then called on a training set with 
   * different structure, and then again with the original training set. 
   * If the equals method of the ASEvaluation class returns false, this is 
   * noted as incorrect search initialisation.
   *
   * @param nominalPredictor if true use nominal predictor attributes
   * @param numericPredictor if true use numeric predictor attributes
   * @param stringPredictor if true use string predictor attributes
   * @param datePredictor if true use date predictor attributes
   * @param relationalPredictor if true use relational predictor attributes
   * @param multiInstance whether multi-instance is needed
   * @param classType the class type (NUMERIC, NOMINAL, etc.)
   * @return index 0 is true if the test was passed, index 1 is always false
   */
  protected boolean[] correctSearchInitialisation(
      boolean nominalPredictor,
      boolean numericPredictor, 
      boolean stringPredictor, 
      boolean datePredictor,
      boolean relationalPredictor,
      boolean multiInstance,
      int classType) {

    boolean[] result = new boolean[2];
    
    print("correct initialisation during search");
    printAttributeSummary(
        nominalPredictor, numericPredictor, stringPredictor, datePredictor, relationalPredictor, multiInstance, classType);
    print("...");
    int numTrain = getNumInstances(), 
    numClasses = 2, missingLevel = 0;
    boolean predictorMissing = false, classMissing = false;
    
    Instances train1 = null;
    Instances train2 = null;
    ASSearch search = null;
    ASEvaluation evaluation1A = null;
    ASEvaluation evaluation1B = null;
    ASEvaluation evaluation2 = null;
    AttributeSelection attsel1A = null;
    AttributeSelection attsel1B = null;
    int stage = 0;
    try {
      
      // Make two train sets with different numbers of attributes
      train1 = makeTestDataset(42, numTrain, 
                               nominalPredictor    ? getNumNominal()    : 0,
                               numericPredictor    ? getNumNumeric()    : 0, 
                               stringPredictor     ? getNumString()     : 0, 
                               datePredictor       ? getNumDate()       : 0, 
                               relationalPredictor ? getNumRelational() : 0, 
                               numClasses, 
                               classType,
                               multiInstance);
      train2 = makeTestDataset(84, numTrain, 
                               nominalPredictor    ? getNumNominal() + 1 : 0,
                               numericPredictor    ? getNumNumeric() + 1 : 0, 
                               stringPredictor     ? getNumString()      : 0, 
                               datePredictor       ? getNumDate()        : 0, 
                               relationalPredictor ? getNumRelational()  : 0, 
                               numClasses, 
                               classType,
                               multiInstance);
      if (missingLevel > 0) {
        addMissing(train1, missingLevel, predictorMissing, classMissing);
        addMissing(train2, missingLevel, predictorMissing, classMissing);
      }
      
      search = ASSearch.makeCopies(getSearch(), 1)[0];
      evaluation1A = ASEvaluation.makeCopies(getEvaluator(), 1)[0];
      evaluation1B = ASEvaluation.makeCopies(getEvaluator(), 1)[0];
      evaluation2 = ASEvaluation.makeCopies(getEvaluator(), 1)[0];
    } catch (Exception ex) {
      throw new Error("Error setting up for tests: " + ex.getMessage());
    }
    try {
      stage = 0;
      attsel1A = search(search, evaluation1A, train1);
      
      stage = 1;
      search(search, evaluation2, train2);
      
      stage = 2;
      attsel1B = search(search, evaluation1B, train1);
      
      stage = 3;
      if (!attsel1A.toResultsString().equals(attsel1B.toResultsString())) {
        if (m_Debug) {
          println(
              "\n=== Full report ===\n"
              + "\nFirst search\n"
              + evaluation1A.toString()
              + "\n\n");
          println(
              "\nSecond search\n"
              + evaluation1B.toString()
              + "\n\n");
        }
        throw new Exception("Results differ between search calls");
      }
      println("yes");
      result[0] = true;
      
      if (false && m_Debug) {
        println(
            "\n=== Full report ===\n"
            + "\nFirst search\n"
            + evaluation1A.toString()
            + "\n\n");
        println(
            "\nSecond search\n"
            + evaluation1B.toString()
            + "\n\n");
      }
    } 
    catch (Exception ex) {
      println("no");
      result[0] = false;
      if (m_Debug) {
        println("\n=== Full Report ===");
        print("Problem during  training");
        switch (stage) {
          case 0:
            print(" of dataset 1");
            break;
          case 1:
            print(" of dataset 2");
            break;
          case 2:
            print(" of dataset 1 (2nd build)");
            break;
          case 3:
            print(", comparing results from builds of dataset 1");
            break;	  
        }
        println(": " + ex.getMessage() + "\n");
        println("here are the datasets:\n");
        println("=== Train1 Dataset ===\n"
            + train1.toString() + "\n");
        println("=== Train2 Dataset ===\n"
            + train2.toString() + "\n");
      }
    }
    
    return result;
  }
  
  /**
   * Checks basic missing value handling of the scheme. If the missing
   * values cause an exception to be thrown by the scheme, this will be
   * recorded.
   *
   * @param nominalPredictor if true use nominal predictor attributes
   * @param numericPredictor if true use numeric predictor attributes
   * @param stringPredictor if true use string predictor attributes
   * @param datePredictor if true use date predictor attributes
   * @param relationalPredictor if true use relational predictor attributes
   * @param multiInstance whether multi-instance is needed
   * @param classType the class type (NUMERIC, NOMINAL, etc.)
   * @param predictorMissing true if the missing values may be in 
   * the predictors
   * @param classMissing true if the missing values may be in the class
   * @param missingLevel the percentage of missing values
   * @return index 0 is true if the test was passed, index 1 is true if test 
   *         was acceptable
   */
  protected boolean[] canHandleMissing(
      boolean nominalPredictor,
      boolean numericPredictor, 
      boolean stringPredictor, 
      boolean datePredictor,
      boolean relationalPredictor,
      boolean multiInstance,
      int classType,
      boolean predictorMissing,
      boolean classMissing,
      int missingLevel) {
    
    if (missingLevel == 100)
      print("100% ");
    print("missing");
    if (predictorMissing) {
      print(" predictor");
      if (classMissing)
        print(" and");
    }
    if (classMissing)
      print(" class");
    print(" values");
    printAttributeSummary(
        nominalPredictor, numericPredictor, stringPredictor, datePredictor, relationalPredictor, multiInstance, classType);
    print("...");
    FastVector accepts = new FastVector();
    accepts.addElement("missing");
    accepts.addElement("value");
    accepts.addElement("train");
    accepts.addElement("no attributes");
    int numTrain = getNumInstances(), numClasses = 2;
    
    return runBasicTest(nominalPredictor, numericPredictor, stringPredictor, 
        datePredictor, relationalPredictor, 
        multiInstance,
        classType, 
        missingLevel, predictorMissing, classMissing,
        numTrain, numClasses, 
        accepts);
  }
  
  /**
   * Checks whether the scheme can handle instance weights.
   * This test compares the scheme performance on two datasets
   * that are identical except for the training weights. If the 
   * results change, then the scheme must be using the weights. It
   * may be possible to get a false positive from this test if the 
   * weight changes aren't significant enough to induce a change
   * in scheme performance (but the weights are chosen to minimize
   * the likelihood of this).
   *
   * @param nominalPredictor if true use nominal predictor attributes
   * @param numericPredictor if true use numeric predictor attributes
   * @param stringPredictor if true use string predictor attributes
   * @param datePredictor if true use date predictor attributes
   * @param relationalPredictor if true use relational predictor attributes
   * @param multiInstance whether multi-instance is needed
   * @param classType the class type (NUMERIC, NOMINAL, etc.)
   * @return index 0 true if the test was passed
   */
  protected boolean[] instanceWeights(
      boolean nominalPredictor,
      boolean numericPredictor, 
      boolean stringPredictor, 
      boolean datePredictor,
      boolean relationalPredictor,
      boolean multiInstance,
      int classType) {
    
    print("scheme uses instance weights");
    printAttributeSummary(
        nominalPredictor, numericPredictor, stringPredictor, datePredictor, relationalPredictor, multiInstance, classType);
    print("...");
    int numTrain = 2*getNumInstances(), 
    numClasses = 2, missingLevel = 0;
    boolean predictorMissing = false, classMissing = false;
    
    boolean[] result = new boolean[2];
    Instances train = null;
    ASSearch[] search = null;
    ASEvaluation evaluationB = null;
    ASEvaluation evaluationI = null;
    AttributeSelection attselB = null;
    AttributeSelection attselI = null;
    boolean evalFail = false;
    try {
      train = makeTestDataset(42, numTrain, 
                              nominalPredictor    ? getNumNominal() + 1 : 0,
                              numericPredictor    ? getNumNumeric() + 1 : 0, 
                              stringPredictor     ? getNumString()      : 0, 
                              datePredictor       ? getNumDate()        : 0, 
                              relationalPredictor ? getNumRelational()  : 0, 
                              numClasses, 
                              classType,
                              multiInstance);
      if (missingLevel > 0)
        addMissing(train, missingLevel, predictorMissing, classMissing);
      search = ASSearch.makeCopies(getSearch(), 2);
      evaluationB = ASEvaluation.makeCopies(getEvaluator(), 1)[0];
      evaluationI = ASEvaluation.makeCopies(getEvaluator(), 1)[0];
      attselB = search(search[0], evaluationB, train);
    } catch (Exception ex) {
      throw new Error("Error setting up for tests: " + ex.getMessage());
    }
    try {
      
      // Now modify instance weights and re-built/test
      for (int i = 0; i < train.numInstances(); i++) {
        train.instance(i).setWeight(0);
      }
      Random random = new Random(1);
      for (int i = 0; i < train.numInstances() / 2; i++) {
        int inst = Math.abs(random.nextInt()) % train.numInstances();
        int weight = Math.abs(random.nextInt()) % 10 + 1;
        train.instance(inst).setWeight(weight);
      }
      attselI = search(search[1], evaluationI, train);
      if (attselB.toResultsString().equals(attselI.toResultsString())) {
        //	println("no");
        evalFail = true;
        throw new Exception("evalFail");
      }
      
      println("yes");
      result[0] = true;
    } catch (Exception ex) {
      println("no");
      result[0] = false;
      
      if (m_Debug) {
        println("\n=== Full Report ===");
        
        if (evalFail) {
          println("Results don't differ between non-weighted and "
              + "weighted instance models.");
          println("Here are the results:\n");
          println("\nboth methods\n");
          println(evaluationB.toString());
        } else {
          print("Problem during training");
          println(": " + ex.getMessage() + "\n");
        }
        println("Here is the dataset:\n");
        println("=== Train Dataset ===\n"
            + train.toString() + "\n");
        println("=== Train Weights ===\n");
        for (int i = 0; i < train.numInstances(); i++) {
          println(" " + (i + 1) 
              + "    " + train.instance(i).weight());
        }
      }
    }
    
    return result;
  }
  
  /**
   * Checks whether the scheme alters the training dataset during
   * training. If the scheme needs to modify the training
   * data it should take a copy of the training data. Currently checks
   * for changes to header structure, number of instances, order of