• Home
• nz.ac.waikato.cms.weka
• isolationForest
• 1.0.2
• source code
• IsolationForest.java

×

Project download

Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance.
Project price only 1 $

You can buy this project and download/modify it how often you want.

weka.classifiers.misc.IsolationForest Maven / Gradle / Ivy

/*
 *   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 3 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, see .
 */

/*
 *    IsolationForest.java
 *    Copyright (C) 2012-16 University of Waikato, Hamilton, New Zealand
 *
 */
package weka.classifiers.misc;

import java.io.Serializable;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Enumeration;
import java.util.Random;
import java.util.Vector;

import weka.classifiers.RandomizableClassifier;
import weka.core.Capabilities;
import weka.core.Capabilities.Capability;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.Option;
import weka.core.TechnicalInformation;
import weka.core.TechnicalInformation.Field;
import weka.core.TechnicalInformation.Type;
import weka.core.TechnicalInformationHandler;
import weka.core.Utils;

/**
 * 
 * Implements the isolation forest method for anomaly detection.

 * 

 * Note that this classifier is designed for anomaly detection, it is not designed for solving two-class or multi-class classification problems!

 * 

 * The data is expected to have have a class attribute with one or two values, which is ignored at training time. The distributionForInstance() method returns (1 - anomaly score) as the first element in the distribution, the second element (in the case of two classes) is the anomaly score.

 * 

 * To evaluate performance of this method for a dataset where anomalies are known, simply code the anomalies using the class attribute: normal cases should correspond to the first value of the class attribute, anomalies to the second one.

 * 

 * For more information, see:

 * 

 * Fei Tony Liu, Kai Ming Ting, Zhi-Hua Zhou: Isolation Forest. In: ICDM, 413-422, 2008.
 * 


 * 
 * 
 * 
 * BibTeX:
 * 
 * @inproceedings{Liu2008,
 *    author = {Fei Tony Liu and Kai Ming Ting and Zhi-Hua Zhou},
 *    booktitle = {ICDM},
 *    pages = {413-422},
 *    publisher = {IEEE Computer Society},
 *    title = {Isolation Forest},
 *    year = {2008}
 * }
 * 
 * 


 * 
 * 
 * 
 * Valid options are: 

 * 
 * 
 -I <number of trees>
 *  The number of trees in the forest (default 100).
 * 
 * 
 -N <the size of the subsample for each tree>
 *  The subsample size for each tree (default 256).
 * 
 * 
 -S <num>
 *  Random number seed.
 *  (default 1)
 * 
 * 
 -output-debug-info
 *  If set, classifier is run in debug mode and
 *  may output additional info to the console
 * 
 * 
 -do-not-check-capabilities
 *  If set, classifier capabilities are not checked before classifier is built
 *  (use with caution).
 * 
 * 
 -num-decimal-places
 *  The number of decimal places for the output of numbers in the model (default 2).
 * 
 * 
 * 
 * @author Eibe Frank ([email protected])
 * @version $Revision: 12345 $
 */
public class IsolationForest extends RandomizableClassifier implements
  TechnicalInformationHandler, Serializable {

  // For serialization
  private static final long serialVersionUID = 5586674623147772788L;

  // The set of trees
  protected Tree[] m_trees = null;

  // The number of trees
  protected int m_numTrees = 100;

  // The subsample size
  protected int m_subsampleSize = 256;

  /**
   * Returns a string describing this filter
   */
  public String globalInfo() {

    return "Implements the isolation forest method for anomaly detection.\n\n"
      + "Note that this classifier is designed for anomaly detection, it is not designed for solving " 
      + "two-class or multi-class classification problems!\n\n"
      + "The data is expected to have have a class attribute with one or two values, "
      + "which is ignored at training time. The distributionForInstance() "
      + "method returns (1 - anomaly score) as the first element in the distribution, "
      + "the second element (in the case of two classes) is the anomaly score.\n\nTo evaluate performance "
      + "of this method for a dataset where anomalies are known, simply "
      + "code the anomalies using the class attribute: normal cases should "
      + "correspond to the first value of the class attribute, anomalies to "
      + "the second one." + "\n\nFor more information, see:\n\n"
      + getTechnicalInformation().toString();
  }

  /**
   * Returns an instance of a TechnicalInformation object, containing detailed
   * information about the technical background of this class, e.g., paper
   * reference or book this class is based on.
   * 
   * @return the technical information about this class
   */
  @Override
  public TechnicalInformation getTechnicalInformation() {
    TechnicalInformation result;

    result = new TechnicalInformation(Type.INPROCEEDINGS);
    result.setValue(Field.AUTHOR,
      "Fei Tony Liu and Kai Ming Ting and Zhi-Hua Zhou");
    result.setValue(Field.TITLE, "Isolation Forest");
    result.setValue(Field.BOOKTITLE, "ICDM");
    result.setValue(Field.YEAR, "2008");
    result.setValue(Field.PAGES, "413-422");
    result.setValue(Field.PUBLISHER, "IEEE Computer Society");

    return result;
  }

  /**
   * Returns the Capabilities of this filter.
   */
  @Override
  public Capabilities getCapabilities() {
    Capabilities result = super.getCapabilities();
    result.disableAll();

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

    // class
    result.enable(Capability.UNARY_CLASS);
    result.enable(Capability.BINARY_CLASS);
    result.enable(Capability.MISSING_CLASS_VALUES);

    // instances
    result.setMinimumNumberInstances(0);

    return result;
  }

  /**
   * Returns brief description of the classifier.
   */
  @Override
  public String toString() {

    if (m_trees == null) {
      return "No model built yet.";
    } else {
      return "Isolation forest for anomaly detection (" + m_numTrees + ", "
        + m_subsampleSize + ")";
    }
  }

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

    return "The number of trees to use in the forest.";
  }

  /**
   * Get the value of numTrees.
   * 
   * @return Value of numTrees.
   */
  public int getNumTrees() {

    return m_numTrees;
  }

  /**
   * Set the value of numTrees.
   * 
   * @param k value to assign to numTrees.
   */
  public void setNumTrees(int k) {

    m_numTrees = k;
  }

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

    return "The size of the subsample used to build each tree.";
  }

  /**
   * Get the value of subsampleSize.
   * 
   * @return Value of subsampleSize.
   */
  public int getSubsampleSize() {

    return m_subsampleSize;
  }

  /**
   * Set the value of subsampleSize.
   * 
   * @param n value to assign to subsampleSize.
   */
  public void setSubsampleSize(int n) {

    m_subsampleSize = n;
  }

  /**
   * Lists the command-line options for this classifier.
   * 
   * @return an enumeration over all possible options
   */
  @Override
  public Enumeration listOptions() {

    Vector newVector = new Vector();

    newVector.addElement(new Option(
      "\tThe number of trees in the forest (default 100).", "I", 1,
      "-I "));

    newVector.addElement(new Option(
      "\tThe subsample size for each tree (default 256).", "N", 1,
      "-N "));

    newVector.addAll(Collections.list(super.listOptions()));

    return newVector.elements();
  }

  /**
   * Gets options from this classifier.
   * 
   * @return the options for the current setup
   */
  @Override
  public String[] getOptions() {

    Vector result = new Vector();

    result.add("-I");
    result.add("" + getNumTrees());

    result.add("-N");
    result.add("" + getSubsampleSize());

    Collections.addAll(result, super.getOptions());

    return result.toArray(new String[result.size()]);
  }

  /**
   * Parses a given list of options.
   * 

   * 
   * 
   * Valid options are: 

   * 
   * 
 -I <number of trees>
   *  The number of trees in the forest (default 100).
   * 
   * 
 -N <the size of the subsample for each tree>
   *  The subsample size for each tree (default 256).
   * 
   * 
 -S <num>
   *  Random number seed.
   *  (default 1)
   * 
   * 
 -output-debug-info
   *  If set, classifier is run in debug mode and
   *  may output additional info to the console
   * 
   * 
 -do-not-check-capabilities
   *  If set, classifier capabilities are not checked before classifier is built
   *  (use with caution).
   * 
   * 
 -num-decimal-places
   *  The number of decimal places for the output of numbers in the model (default 2).
   * 
   * 
   * 
   * @param options the list of options as an array of strings
   * @throws Exception if an option is not supported
   */
  @Override
  public void setOptions(String[] options) throws Exception {
    String tmpStr;

    tmpStr = Utils.getOption('I', options);
    if (tmpStr.length() != 0) {
      m_numTrees = Integer.parseInt(tmpStr);
    } else {
      m_numTrees = 100;
    }

    tmpStr = Utils.getOption('N', options);
    if (tmpStr.length() != 0) {
      m_subsampleSize = Integer.parseInt(tmpStr);
    } else {
      m_subsampleSize = 256;
    }

    super.setOptions(options);

    Utils.checkForRemainingOptions(options);
  }

  /**
   * Builds the forest.
   */
  @Override
  public void buildClassifier(Instances data) throws Exception {

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

    // Reduce subsample size if data is too small
    if (data.numInstances() < m_subsampleSize) {
      m_subsampleSize = data.numInstances();
    }

    // Generate trees
    m_trees = new Tree[m_numTrees];
    data = new Instances(data);
    Random r = (data.numInstances() > 0) ? data
      .getRandomNumberGenerator(m_Seed) : new Random(m_Seed);
    for (int i = 0; i < m_numTrees; i++) {
      data.randomize(r);
      m_trees[i] = new Tree(new Instances(data, 0, m_subsampleSize), r, 0,
        (int) Math.ceil(Utils.log2(data.numInstances())));
    }
  }

  /**
   * Returns the average path length of an unsuccessful search. Returns 0 if
   * argument is less than or equal to 1
   */
  public static double c(double n) {

    if (n <= 1.0) {
      return 0;
    }
    return 2 * (Math.log(n - 1) + 0.5772156649) - (2 * (n - 1) / n);
  }

  /**
   * Returns distribution of scores.
   */
  @Override
  public double[] distributionForInstance(Instance inst) {

    double avgPathLength = 0;
    for (Tree m_tree : m_trees) {
      avgPathLength += m_tree.pathLength(inst);
    }
    avgPathLength /= m_trees.length;

    double[] scores = new double[inst.numClasses()];
    scores[0] = 1.0 - Math.pow(2, -avgPathLength / c(m_subsampleSize));
    if (scores.length > 1) {
      scores[1] = 1.0 - scores[0];
    }

    return scores;
  }

  /**
   * Main method for this class.
   */
  public static void main(String[] args) {

    runClassifier(new IsolationForest(), args);
  }

  /**
   * Inner class for building and using an isolation tree.
   */
  protected class Tree implements Serializable {

    // For serialization
    private static final long serialVersionUID = 7786674623147772711L;

    // The size of the node
    protected int m_size;

    // The split attribute
    protected int m_a;

    // The split point
    protected double m_splitPoint;

    // The successors
    protected Tree[] m_successors;

    /**
     * Constructs a tree from data
     */
    protected Tree(Instances data, Random r, int height, int maxHeight) {

      // Set size of node
      m_size = data.numInstances();

      // Stop splitting if necessary
      if ((m_size <= 1) || (height == maxHeight)) {
        return;
      }

      // Compute mins and maxs and eligible attributes
      ArrayList al = new ArrayList();
      double[][] minmax = new double[2][data.numAttributes()];
      for (int j = 0; j < data.numAttributes(); j++) {
        minmax[0][j] = data.instance(0).value(j);
        minmax[1][j] = minmax[0][j];
      }
      for (int i = 1; i < data.numInstances(); i++) {
        Instance inst = data.instance(i);
        for (int j = 0; j < data.numAttributes(); j++) {
          if (inst.value(j) < minmax[0][j]) {
            minmax[0][j] = inst.value(j);
          }
          if (inst.value(j) > minmax[1][j]) {
            minmax[1][j] = inst.value(j);
          }
        }
      }
      for (int j = 0; j < data.numAttributes(); j++) {
        if (j != data.classIndex()) {
          if (minmax[0][j] < minmax[1][j]) {
            al.add(j);
          }
        }
      }

      // Check whether any eligible attributes have been found
      if (al.size() == 0) {
        return;
      } else {

        // Randomly pick an attribute and split point
        m_a = al.get(r.nextInt(al.size()));
        m_splitPoint = (r.nextDouble() * (minmax[1][m_a] - minmax[0][m_a]))
          + minmax[0][m_a];

        // Create sub trees
        m_successors = new Tree[2];
        for (int i = 0; i < 2; i++) {
          Instances tempData = new Instances(data, data.numInstances());
          for (int j = 0; j < data.numInstances(); j++) {
            if ((i == 0) && (data.instance(j).value(m_a) < m_splitPoint)) {
              tempData.add(data.instance(j));
            }
            if ((i == 1) && (data.instance(j).value(m_a) >= m_splitPoint)) {
              tempData.add(data.instance(j));
            }
          }
          tempData.compactify();
          m_successors[i] = new Tree(tempData, r, height + 1, maxHeight);
        }
      }
    }

    /**
     * Returns path length according to algorithm.
     */
    protected double pathLength(Instance inst) {

      if (m_successors == null) {
        return c(m_size);
      }
      if (inst.value(m_a) < m_splitPoint) {
        return m_successors[0].pathLength(inst) + 1.0;
      } else {
        return m_successors[1].pathLength(inst) + 1.0;
      }
    }
  }
}