ibk.java

来自「Weka」· Java 代码 · 共 1,028 行 · 第 1/3 页

   * Returns the tip text for this property.
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String meanSquaredTipText() {

    return "Whether the mean squared error is used rather than mean "
      + "absolute error when doing cross-validation for regression problems.";
  }

  /**
   * Gets whether the mean squared error is used rather than mean
   * absolute error when doing cross-validation.
   *
   * @return true if so.
   */
  public boolean getMeanSquared() {
    
    return m_MeanSquared;
  }
  
  /**
   * Sets whether the mean squared error is used rather than mean
   * absolute error when doing cross-validation.
   *
   * @param newMeanSquared true if so.
   */
  public void setMeanSquared(boolean newMeanSquared) {
    
    m_MeanSquared = newMeanSquared;
  }
  
  /**
   * Returns the tip text for this property.
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String crossValidateTipText() {

    return "Whether hold-one-out cross-validation will be used " +
      "to select the best k value.";
  }
  
  /**
   * Gets whether hold-one-out cross-validation will be used
   * to select the best k value.
   *
   * @return true if cross-validation will be used.
   */
  public boolean getCrossValidate() {
    
    return m_CrossValidate;
  }
  
  /**
   * Sets whether hold-one-out cross-validation will be used
   * to select the best k value.
   *
   * @param newCrossValidate true if cross-validation should be used.
   */
  public void setCrossValidate(boolean newCrossValidate) {
    
    m_CrossValidate = newCrossValidate;
  }

  /**
   * Returns the tip text for this property.
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String nearestNeighbourSearchAlgorithmTipText() {
    return "The nearest neighbour search algorithm to use " +
    	   "(Default: weka.core.neighboursearch.LinearNNSearch).";
  }
  
  /**
   * Returns the current nearestNeighbourSearch algorithm in use.
   * @return the NearestNeighbourSearch algorithm currently in use.
   */
  public NearestNeighbourSearch getNearestNeighbourSearchAlgorithm() {
    return m_NNSearch;
  }
  
  /**
   * Sets the nearestNeighbourSearch algorithm to be used for finding nearest
   * neighbour(s).
   * @param nearestNeighbourSearchAlgorithm - The NearestNeighbourSearch class.
   */
  public void setNearestNeighbourSearchAlgorithm(NearestNeighbourSearch nearestNeighbourSearchAlgorithm) {
    m_NNSearch = nearestNeighbourSearchAlgorithm;
  }
   
  /**
   * Get the number of training instances the classifier is currently using.
   * 
   * @return the number of training instances the classifier is currently using
   */
  public int getNumTraining() {

    return m_Train.numInstances();
  }

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

    // attributes
    result.enable(Capability.NOMINAL_ATTRIBUTES);
    result.enable(Capability.NUMERIC_ATTRIBUTES);
    result.enable(Capability.DATE_ATTRIBUTES);
    result.enable(Capability.MISSING_VALUES);

    // class
    result.enable(Capability.NOMINAL_CLASS);
    result.enable(Capability.NUMERIC_CLASS);
    result.enable(Capability.DATE_CLASS);
    result.enable(Capability.MISSING_CLASS_VALUES);

    // instances
    result.setMinimumNumberInstances(0);
    
    return result;
  }
  
  /**
   * Generates the classifier.
   *
   * @param instances set of instances serving as training data 
   * @throws Exception if the classifier has not been generated successfully
   */
  public void buildClassifier(Instances instances) throws Exception {
    
    // can classifier handle the data?
    getCapabilities().testWithFail(instances);

    // remove instances with missing class
    instances = new Instances(instances);
    instances.deleteWithMissingClass();
    
    m_NumClasses = instances.numClasses();
    m_ClassType = instances.classAttribute().type();
    m_Train = new Instances(instances, 0, instances.numInstances());

    // Throw away initial instances until within the specified window size
    if ((m_WindowSize > 0) && (instances.numInstances() > m_WindowSize)) {
      m_Train = new Instances(m_Train, 
			      m_Train.numInstances()-m_WindowSize, 
			      m_WindowSize);
    }

    m_NumAttributesUsed = 0.0;
    for (int i = 0; i < m_Train.numAttributes(); i++) {
      if ((i != m_Train.classIndex()) && 
	  (m_Train.attribute(i).isNominal() ||
	   m_Train.attribute(i).isNumeric())) {
	m_NumAttributesUsed += 1.0;
      }
    }
    
    m_NNSearch.setInstances(m_Train);

    // Invalidate any currently cross-validation selected k
    m_kNNValid = false;
  }

  /**
   * Adds the supplied instance to the training set.
   *
   * @param instance the instance to add
   * @throws Exception if instance could not be incorporated
   * successfully
   */
  public void updateClassifier(Instance instance) throws Exception {

    if (m_Train.equalHeaders(instance.dataset()) == false) {
      throw new Exception("Incompatible instance types");
    }
    if (instance.classIsMissing()) {
      return;
    }

    m_Train.add(instance);
    m_NNSearch.update(instance);
    m_kNNValid = false;
    if ((m_WindowSize > 0) && (m_Train.numInstances() > m_WindowSize)) {
      boolean deletedInstance=false;
      while (m_Train.numInstances() > m_WindowSize) {
	m_Train.delete(0);
        deletedInstance=true;
      }
      //rebuild datastructure KDTree currently can't delete
      if(deletedInstance==true)
        m_NNSearch.setInstances(m_Train);
    }
  }

  /**
   * Calculates the class membership probabilities for the given test instance.
   *
   * @param instance the instance to be classified
   * @return predicted class probability distribution
   * @throws Exception if an error occurred during the prediction
   */
  public double [] distributionForInstance(Instance instance) throws Exception {

    if (m_Train.numInstances() == 0) {
      throw new Exception("No training instances!");
    }
    if ((m_WindowSize > 0) && (m_Train.numInstances() > m_WindowSize)) {
      m_kNNValid = false;
      boolean deletedInstance=false;
      while (m_Train.numInstances() > m_WindowSize) {
	m_Train.delete(0);
      }
      //rebuild datastructure KDTree currently can't delete
      if(deletedInstance==true)
        m_NNSearch.setInstances(m_Train);
    }

    // Select k by cross validation
    if (!m_kNNValid && (m_CrossValidate) && (m_kNNUpper >= 1)) {
      crossValidate();
    }

    m_NNSearch.addInstanceInfo(instance);

    Instances neighbours = m_NNSearch.kNearestNeighbours(instance, m_kNN);
    double [] distances = m_NNSearch.getDistances();
    double [] distribution = makeDistribution( neighbours, distances );

    return distribution;
  }

  /**
   * Returns an enumeration describing the available options.
   *
   * @return an enumeration of all the available options.
   */
  public Enumeration listOptions() {

    Vector newVector = new Vector(8);

    newVector.addElement(new Option(
	      "\tWeight neighbours by the inverse of their distance\n"+
	      "\t(use when k > 1)",
	      "I", 0, "-I"));
    newVector.addElement(new Option(
	      "\tWeight neighbours by 1 - their distance\n"+
	      "\t(use when k > 1)",
	      "F", 0, "-F"));
    newVector.addElement(new Option(
	      "\tNumber of nearest neighbours (k) used in classification.\n"+
	      "\t(Default = 1)",
	      "K", 1,"-K <number of neighbors>"));
    newVector.addElement(new Option(
          "\tMinimise mean squared error rather than mean absolute\n"+
	      "\terror when using -X option with numeric prediction.",
	      "E", 0,"-E"));
    newVector.addElement(new Option(
          "\tMaximum number of training instances maintained.\n"+
	      "\tTraining instances are dropped FIFO. (Default = no window)",
	      "W", 1,"-W <window size>"));
    newVector.addElement(new Option(
	      "\tSelect the number of nearest neighbours between 1\n"+
	      "\tand the k value specified using hold-one-out evaluation\n"+
	      "\ton the training data (use when k > 1)",
	      "X", 0,"-X"));
    newVector.addElement(new Option(
	      "\tThe nearest neighbour search algorithm to use "+
          "(default: weka.core.neighboursearch.LinearNNSearch).\n",
	      "A", 0, "-A"));

    return newVector.elements();
  }

  /**
   * Parses a given list of options. <p/>
   *
   <!-- options-start -->
   * Valid options are: <p/>
   * 
   * <pre> -I
   *  Weight neighbours by the inverse of their distance
   *  (use when k &gt; 1)</pre>
   * 
   * <pre> -F
   *  Weight neighbours by 1 - their distance
   *  (use when k &gt; 1)</pre>
   * 
   * <pre> -K &lt;number of neighbors&gt;
   *  Number of nearest neighbours (k) used in classification.
   *  (Default = 1)</pre>
   * 
   * <pre> -E
   *  Minimise mean squared error rather than mean absolute
   *  error when using -X option with numeric prediction.</pre>
   * 
   * <pre> -W &lt;window size&gt;
   *  Maximum number of training instances maintained.
   *  Training instances are dropped FIFO. (Default = no window)</pre>
   * 
   * <pre> -X
   *  Select the number of nearest neighbours between 1
   *  and the k value specified using hold-one-out evaluation
   *  on the training data (use when k &gt; 1)</pre>
   * 
   * <pre> -A
   *  The nearest neighbour search algorithm to use (default: weka.core.neighboursearch.LinearNNSearch).
   * </pre>
   * 
   <!-- options-end -->
   *
   * @param options the list of options as an array of strings
   * @throws Exception if an option is not supported
   */
  public void setOptions(String[] options) throws Exception {
    
    String knnString = Utils.getOption('K', options);
    if (knnString.length() != 0) {
      setKNN(Integer.parseInt(knnString));
    } else {
      setKNN(1);
    }
    String windowString = Utils.getOption('W', options);
    if (windowString.length() != 0) {
      setWindowSize(Integer.parseInt(windowString));
    } else {
      setWindowSize(0);
    }
    if (Utils.getFlag('I', options)) {
      setDistanceWeighting(new SelectedTag(WEIGHT_INVERSE, TAGS_WEIGHTING));
    } else if (Utils.getFlag('F', options)) {
      setDistanceWeighting(new SelectedTag(WEIGHT_SIMILARITY, TAGS_WEIGHTING));
    } else {
      setDistanceWeighting(new SelectedTag(WEIGHT_NONE, TAGS_WEIGHTING));
    }
    setCrossValidate(Utils.getFlag('X', options));
    setMeanSquared(Utils.getFlag('E', options));