classbalancednd.java

代码是一个分类器的实现,其中使用了部分weka的源代码。可以将项目导入eclipse运行

      firstCopy[i] = firstInds[sortedFirst[i]];
    }
    firstInds = firstCopy;
    for (int i = 0; i < sortedSecond.length; i++) {
      secondCopy[i] = secondInds[sortedSecond[i]];
    }
    secondInds = secondCopy;
		
    // Unify indices to improve hashing
    if (firstInds[0] > secondInds[0]) {
      int[] help = secondInds;
      secondInds = firstInds;
      firstInds = help;
      int help2 = second;
      second = first;
      first = help2;
    }

    m_Range = new Range(Range.indicesToRangeList(firstInds));
    m_Range.setUpper(data.numClasses() - 1);

    Range secondRange = new Range(Range.indicesToRangeList(secondInds));
    secondRange.setUpper(data.numClasses() - 1);
       
    // Change the class labels and build the classifier
    MakeIndicator filter = new MakeIndicator();
    filter.setAttributeIndex("" + (data.classIndex() + 1));
    filter.setValueIndices(m_Range.getRanges());
    filter.setNumeric(false);
    filter.setInputFormat(data);
    m_FilteredClassifier = new FilteredClassifier();
    if (data.numInstances() > 0) {
      m_FilteredClassifier.setClassifier(Classifier.makeCopies(classifier, 1)[0]);
    } else {
      m_FilteredClassifier.setClassifier(new weka.classifiers.rules.ZeroR());
    }
    m_FilteredClassifier.setFilter(filter);

    // Save reference to hash table at current node
    m_classifiers=table;
	
    if (!m_classifiers.containsKey( getString(firstInds) + "|" + getString(secondInds))) {
      m_FilteredClassifier.buildClassifier(data);
      m_classifiers.put(getString(firstInds) + "|" + getString(secondInds), m_FilteredClassifier);
    } else {
      m_FilteredClassifier=(FilteredClassifier)m_classifiers.get(getString(firstInds) + "|" + 
								 getString(secondInds));	
    }
				
    // Create two successors if necessary
    m_FirstSuccessor = new ClassBalancedND();
    if (first == 1) {
      m_FirstSuccessor.m_Range = m_Range;
    } else {
      RemoveWithValues rwv = new RemoveWithValues();
      rwv.setInvertSelection(true);
      rwv.setNominalIndices(m_Range.getRanges());
      rwv.setAttributeIndex("" + (data.classIndex() + 1));
      rwv.setInputFormat(data);
      Instances firstSubset = Filter.useFilter(data, rwv);
      m_FirstSuccessor.generateClassifierForNode(firstSubset, m_Range, 
                                                 rand, classifier, m_classifiers);
    }
    m_SecondSuccessor = new ClassBalancedND();
    if (second == 1) {
      m_SecondSuccessor.m_Range = secondRange;
    } else {
      RemoveWithValues rwv = new RemoveWithValues();
      rwv.setInvertSelection(true);
      rwv.setNominalIndices(secondRange.getRanges());
      rwv.setAttributeIndex("" + (data.classIndex() + 1));
      rwv.setInputFormat(data);
      Instances secondSubset = Filter.useFilter(data, rwv);
      m_SecondSuccessor = new ClassBalancedND();
      
      m_SecondSuccessor.generateClassifierForNode(secondSubset, secondRange, 
                                                  rand, classifier, m_classifiers);
    }
  }

  /**
   * Returns default capabilities of the classifier.
   *
   * @return      the capabilities of this classifier
   */
  public Capabilities getCapabilities() {
    Capabilities result = super.getCapabilities();

    // class
    result.disableAllClasses();
    result.enable(Capability.NOMINAL_CLASS);
    result.enable(Capability.MISSING_CLASS_VALUES);

    // instances
    result.setMinimumNumberInstances(1);
    
    return result;
  }
    
  /**
   * Builds tree recursively.
   *
   * @param data contains the (multi-class) instances
   * @throws Exception if the building fails
   */
  public void buildClassifier(Instances data) throws Exception {

    // can classifier handle the data?
    getCapabilities().testWithFail(data);

    // remove instances with missing class
    data = new Instances(data);
    data.deleteWithMissingClass();
    
    Random random = data.getRandomNumberGenerator(m_Seed);
	
    if (!m_hashtablegiven) {
      m_classifiers = new Hashtable();
    }
	
    // Check which classes are present in the
    // data and construct initial list of classes
    boolean[] present = new boolean[data.numClasses()];
    for (int i = 0; i < data.numInstances(); i++) {
      present[(int)data.instance(i).classValue()] = true;
    }
    StringBuffer list = new StringBuffer();
    for (int i = 0; i < present.length; i++) {
      if (present[i]) {
        if (list.length() > 0) {
          list.append(",");
        }
        list.append(i + 1);
      }
    }
      
    Range newRange = new Range(list.toString());
    newRange.setUpper(data.numClasses() - 1);
	
    generateClassifierForNode(data, newRange, random, m_Classifier, m_classifiers);
  }
    
  /**
   * Predicts the class distribution for a given instance
   *
   * @param inst the (multi-class) instance to be classified
   * @return the class distribution
   * @throws Exception if computing fails
   */
  public double[] distributionForInstance(Instance inst) throws Exception {
	
    double[] newDist = new double[inst.numClasses()];
    if (m_FirstSuccessor == null) {
      for (int i = 0; i < inst.numClasses(); i++) {
        if (m_Range.isInRange(i)) {
          newDist[i] = 1;
        }
      }
      return newDist;
    } else {
      double[] firstDist = m_FirstSuccessor.distributionForInstance(inst);
      double[] secondDist = m_SecondSuccessor.distributionForInstance(inst);
      double[] dist = m_FilteredClassifier.distributionForInstance(inst);
      for (int i = 0; i < inst.numClasses(); i++) {
        if ((firstDist[i] > 0) && (secondDist[i] > 0)) {
          System.err.println("Panik!!");
        }
        if (m_Range.isInRange(i)) {
          newDist[i] = dist[1] * firstDist[i];
        } else {
          newDist[i] = dist[0] * secondDist[i];
        }
      }
      return newDist;
    }
  }
    
  /**
   * Returns the list of indices as a string.
   * 
   * @param indices the indices to return as string
   * @return the indices as string
   */
  public String getString(int [] indices) {

    StringBuffer string = new StringBuffer();
    for (int i = 0; i < indices.length; i++) {
      if (i > 0) {
        string.append(',');
      }
      string.append(indices[i]);
    }
    return string.toString();
  }
	
  /**
   * @return a description of the classifier suitable for
   * displaying in the explorer/experimenter gui
   */
  public String globalInfo() {
	    
    return 
        "A meta classifier for handling multi-class datasets with 2-class "
      + "classifiers by building a random class-balanced tree structure.\n\n"
      + "For more info, check\n\n"
      + getTechnicalInformation().toString();
  }
	
  /**
   * Outputs the classifier as a string.
   * 
   * @return a string representation of the classifier
   */
  public String toString() {
	    
    if (m_classifiers == null) {
      return "ClassBalancedND: No model built yet.";
    }
    StringBuffer text = new StringBuffer();
    text.append("ClassBalancedND");
    treeToString(text, 0);
	    
    return text.toString();
  }
	
  /**
   * Returns string description of the tree.
   * 
   * @param text the buffer to add the node to
   * @param nn the node number
   * @return the next node number
   */
  private int treeToString(StringBuffer text, int nn) {
	    
    nn++;
    text.append("\n\nNode number: " + nn + "\n\n");
    if (m_FilteredClassifier != null) {
      text.append(m_FilteredClassifier);
    } else {
      text.append("null");
    }
    if (m_FirstSuccessor != null) {
      nn = m_FirstSuccessor.treeToString(text, nn);
      nn = m_SecondSuccessor.treeToString(text, nn);
    }
    return nn;
  }
    	
  /**
   * Main method for testing this class.
   *
   * @param argv the options
   */
  public static void main(String [] argv) {
    runClassifier(new ClassBalancedND(), argv);
  }
}