类 Decorate

所有已实现的接口:
Serializable, Cloneable, CapabilitiesHandler, OptionHandler, Randomizable, RevisionHandler, TechnicalInformationHandler

DECORATE is a meta-learner for building diverse ensembles of classifiers by using specially constructed artificial training examples. Comprehensive experiments have demonstrated that this technique is consistently more accurate than the base classifier, Bagging and Random Forests.Decorate also obtains higher accuracy than Boosting on small training sets, and achieves comparable performance on larger training sets.

For more details see:

P. Melville, R. J. Mooney: Constructing Diverse Classifier Ensembles Using Artificial Training Examples. In: Eighteenth International Joint Conference on Artificial Intelligence, 505-510, 2003.

P. Melville, R. J. Mooney (2004). Creating Diversity in Ensembles Using Artificial Data. Information Fusion: Special Issue on Diversity in Multiclassifier Systems..

BibTeX:

 @inproceedings{Melville2003,
    author = {P. Melville and R. J. Mooney},
    booktitle = {Eighteenth International Joint Conference on Artificial Intelligence},
    pages = {505-510},
    title = {Constructing Diverse Classifier Ensembles Using Artificial Training Examples},
    year = {2003}
 }
 
 @article{Melville2004,
    author = {P. Melville and R. J. Mooney},
    journal = {Information Fusion: Special Issue on Diversity in Multiclassifier Systems},
    note = {submitted},
    title = {Creating Diversity in Ensembles Using Artificial Data},
    year = {2004}
 }
 

Valid options are:

 -E
  Desired size of ensemble.
  (default 15)
 -R
  Factor that determines number of artificial examples to generate.
  Specified proportional to training set size.
  (default 1.0)
 -S <num>
  Random number seed.
  (default 1)
 -I <num>
  Number of iterations.
  (default 50)
 -D
  If set, classifier is run in debug mode and
  may output additional info to the console
 -W
  Full name of base classifier.
  (default: weka.classifiers.trees.J48)
 
 Options specific to classifier weka.classifiers.trees.J48:
 
 -U
  Use unpruned tree.
 -C <pruning confidence>
  Set confidence threshold for pruning.
  (default 0.25)
 -M <minimum number of instances>
  Set minimum number of instances per leaf.
  (default 2)
 -R
  Use reduced error pruning.
 -N <number of folds>
  Set number of folds for reduced error
  pruning. One fold is used as pruning set.
  (default 3)
 -B
  Use binary splits only.
 -S
  Don't perform subtree raising.
 -L
  Do not clean up after the tree has been built.
 -A
  Laplace smoothing for predicted probabilities.
 -Q <seed>
  Seed for random data shuffling (default 1).
Options after -- are passed to the designated classifier.

版本:
$Revision: 8037 $
作者:
Prem Melville (melville@cs.utexas.edu)
另请参阅:
  • 构造器详细资料

    • Decorate

      public Decorate()
      Constructor.
  • 方法详细资料

    • listOptions

      public Enumeration listOptions()
      Returns an enumeration describing the available options
      指定者:
      listOptions 在接口中 OptionHandler
      覆盖:
      listOptions 在类中 RandomizableIteratedSingleClassifierEnhancer
      返回:
      an enumeration of all the available options
    • setOptions

      public void setOptions(String[] options) throws Exception
      Parses a given list of options.

      Valid options are:

       -E
        Desired size of ensemble.
        (default 15)
       -R
        Factor that determines number of artificial examples to generate.
        Specified proportional to training set size.
        (default 1.0)
       -S <num>
        Random number seed.
        (default 1)
       -I <num>
        Number of iterations.
        (default 50)
       -D
        If set, classifier is run in debug mode and
        may output additional info to the console
       -W
        Full name of base classifier.
        (default: weka.classifiers.trees.J48)
       
       Options specific to classifier weka.classifiers.trees.J48:
       
       -U
        Use unpruned tree.
       -C <pruning confidence>
        Set confidence threshold for pruning.
        (default 0.25)
       -M <minimum number of instances>
        Set minimum number of instances per leaf.
        (default 2)
       -R
        Use reduced error pruning.
       -N <number of folds>
        Set number of folds for reduced error
        pruning. One fold is used as pruning set.
        (default 3)
       -B
        Use binary splits only.
       -S
        Don't perform subtree raising.
       -L
        Do not clean up after the tree has been built.
       -A
        Laplace smoothing for predicted probabilities.
       -Q <seed>
        Seed for random data shuffling (default 1).
      Options after -- are passed to the designated classifier.

      指定者:
      setOptions 在接口中 OptionHandler
      覆盖:
      setOptions 在类中 RandomizableIteratedSingleClassifierEnhancer
      参数:
      options - the list of options as an array of strings
      抛出:
      Exception - if an option is not supported
    • getOptions

      public String[] getOptions()
      Gets the current settings of the Classifier.
      指定者:
      getOptions 在接口中 OptionHandler
      覆盖:
      getOptions 在类中 RandomizableIteratedSingleClassifierEnhancer
      返回:
      an array of strings suitable for passing to setOptions
    • desiredSizeTipText

      public String desiredSizeTipText()
      Returns the tip text for this property
      返回:
      tip text for this property suitable for displaying in the explorer/experimenter gui
    • numIterationsTipText

      public String numIterationsTipText()
      Returns the tip text for this property
      覆盖:
      numIterationsTipText 在类中 IteratedSingleClassifierEnhancer
      返回:
      tip text for this property suitable for displaying in the explorer/experimenter gui
    • artificialSizeTipText

      public String artificialSizeTipText()
      Returns the tip text for this property
      返回:
      tip text for this property suitable for displaying in the explorer/experimenter gui
    • globalInfo

      public String globalInfo()
      Returns a string describing classifier
      返回:
      a description suitable for displaying in the explorer/experimenter gui
    • getTechnicalInformation

      public TechnicalInformation getTechnicalInformation()
      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.
      指定者:
      getTechnicalInformation 在接口中 TechnicalInformationHandler
      返回:
      the technical information about this class
    • getArtificialSize

      public double getArtificialSize()
      Factor that determines number of artificial examples to generate.
      返回:
      factor that determines number of artificial examples to generate
    • setArtificialSize

      public void setArtificialSize(double newArtSize)
      Sets factor that determines number of artificial examples to generate.
      参数:
      newArtSize - factor that determines number of artificial examples to generate
    • getDesiredSize

      public int getDesiredSize()
      Gets the desired size of the committee.
      返回:
      the desired size of the committee
    • setDesiredSize

      public void setDesiredSize(int newDesiredSize)
      Sets the desired size of the committee.
      参数:
      newDesiredSize - the desired size of the committee
    • getCapabilities

      public Capabilities getCapabilities()
      Returns default capabilities of the classifier.
      指定者:
      getCapabilities 在接口中 CapabilitiesHandler
      覆盖:
      getCapabilities 在类中 SingleClassifierEnhancer
      返回:
      the capabilities of this classifier
      另请参阅:
    • buildClassifier

      public void buildClassifier(Instances data) throws Exception
      Build Decorate classifier
      覆盖:
      buildClassifier 在类中 IteratedSingleClassifierEnhancer
      参数:
      data - the training data to be used for generating the classifier
      抛出:
      Exception - if the classifier could not be built successfully
    • distributionForInstance

      public double[] distributionForInstance(Instance instance) throws Exception
      Calculates the class membership probabilities for the given test instance.
      覆盖:
      distributionForInstance 在类中 Classifier
      参数:
      instance - the instance to be classified
      返回:
      predicted class probability distribution
      抛出:
      Exception - if distribution can't be computed successfully
    • toString

      public String toString()
      Returns description of the Decorate classifier.
      覆盖:
      toString 在类中 Object
      返回:
      description of the Decorate classifier as a string
    • getRevision

      public String getRevision()
      Returns the revision string.
      指定者:
      getRevision 在接口中 RevisionHandler
      覆盖:
      getRevision 在类中 Classifier
      返回:
      the revision
    • main

      public static void main(String[] argv)
      Main method for testing this class.
      参数:
      argv - the options