Class SMO

All Implemented Interfaces:
Serializable, Cloneable, Classifier, BatchPredictor, CapabilitiesHandler, CapabilitiesIgnorer, CommandlineRunnable, OptionHandler, RevisionHandler, TechnicalInformationHandler, WeightedInstancesHandler

Implements John Platt's sequential minimal optimization algorithm for training a support vector classifier.

This implementation globally replaces all missing values and transforms nominal attributes into binary ones. It also normalizes all attributes by default. (In that case the coefficients in the output are based on the normalized data, not the original data --- this is important for interpreting the classifier.)

Multi-class problems are solved using pairwise classification (aka 1-vs-1).

To obtain proper probability estimates, use the option that fits calibration models to the outputs of the support vector machine. In the multi-class case, the predicted probabilities are coupled using Hastie and Tibshirani's pairwise coupling method.

Note: for improved speed normalization should be turned off when operating on SparseInstances.

For more information on the SMO algorithm, see

J. Platt: Fast Training of Support Vector Machines using Sequential Minimal Optimization. In B. Schoelkopf and C. Burges and A. Smola, editors, Advances in Kernel Methods - Support Vector Learning, 1998.

S.S. Keerthi, S.K. Shevade, C. Bhattacharyya, K.R.K. Murthy (2001). Improvements to Platt's SMO Algorithm for SVM Classifier Design. Neural Computation. 13(3):637-649.

Trevor Hastie, Robert Tibshirani: Classification by Pairwise Coupling. In: Advances in Neural Information Processing Systems, 1998.

BibTeX:
 @incollection{Platt1998,
    author = {J. Platt},
    booktitle = {Advances in Kernel Methods - Support Vector Learning},
    editor = {B. Schoelkopf and C. Burges and A. Smola},
    publisher = {MIT Press},
    title = {Fast Training of Support Vector Machines using Sequential Minimal Optimization},
    year = {1998},
    URL = {http://research.microsoft.com/\~jplatt/smo.html},
    PS = {http://research.microsoft.com/\~jplatt/smo-book.ps.gz},
    PDF = {http://research.microsoft.com/\~jplatt/smo-book.pdf}
 }
 
 @article{Keerthi2001,
    author = {S.S. Keerthi and S.K. Shevade and C. Bhattacharyya and K.R.K. Murthy},
    journal = {Neural Computation},
    number = {3},
    pages = {637-649},
    title = {Improvements to Platt's SMO Algorithm for SVM Classifier Design},
    volume = {13},
    year = {2001},
    PS = {http://guppy.mpe.nus.edu.sg/\~mpessk/svm/smo_mod_nc.ps.gz}
 }
 
 @inproceedings{Hastie1998,
    author = {Trevor Hastie and Robert Tibshirani},
    booktitle = {Advances in Neural Information Processing Systems},
    editor = {Michael I. Jordan and Michael J. Kearns and Sara A. Solla},
    publisher = {MIT Press},
    title = {Classification by Pairwise Coupling},
    volume = {10},
    year = {1998},
    PS = {http://www-stat.stanford.edu/\~hastie/Papers/2class.ps}
 }
 


Valid options are:

 -no-checks
  Turns off all checks - use with caution!
  Turning them off assumes that data is purely numeric, doesn't
  contain any missing values, and has a nominal class. Turning them
  off also means that no header information will be stored if the
  machine is linear. Finally, it also assumes that no instance has
  a weight equal to 0.
  (default: checks on)
 -C <double>
  The complexity constant C. (default 1)
 -N
  Whether to 0=normalize/1=standardize/2=neither. (default 0=normalize)
 -L <double>
  The tolerance parameter. (default 1.0e-3)
 -P <double>
  The epsilon for round-off error. (default 1.0e-12)
 -M
  Fit calibration models to SVM outputs. 
 -V <double>
  The number of folds for the internal
  cross-validation. (default -1, use training data)
 -W <double>
  The random number seed. (default 1)
 -K <classname and parameters>
  The Kernel to use.
  (default: weka.classifiers.functions.supportVector.PolyKernel)
 -calibrator <scheme specification>
  Full name of calibration model, followed by options.
  (default: "weka.classifiers.functions.Logistic")
 -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).
 
 Options specific to kernel weka.classifiers.functions.supportVector.PolyKernel:
 
 -E <num>
  The Exponent to use.
  (default: 1.0)
 -L
  Use lower-order terms.
  (default: no)
 -C <num>
  The size of the cache (a prime number), 0 for full cache and 
  -1 to turn it off.
  (default: 250007)
 -output-debug-info
  Enables debugging output (if available) to be printed.
  (default: off)
 -no-checks
  Turns off all checks - use with caution!
  (default: checks on)
 
 Options specific to calibrator weka.classifiers.functions.Logistic:
 
 -C
  Use conjugate gradient descent rather than BFGS updates.
 -R <ridge>
  Set the ridge in the log-likelihood.
 -M <number>
  Set the maximum number of iterations (default -1, until convergence).
 -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).
Version:
$Revision: 15561 $
Author:
Eibe Frank (eibe@cs.waikato.ac.nz), Shane Legg (shane@intelligenesis.net) (sparse vector code), Stuart Inglis (stuart@reeltwo.com) (sparse vector code)
See Also:
  • Field Details

    • FILTER_NORMALIZE

      public static final int FILTER_NORMALIZE
      filter: Normalize training data
      See Also:
    • FILTER_STANDARDIZE

      public static final int FILTER_STANDARDIZE
      filter: Standardize training data
      See Also:
    • FILTER_NONE

      public static final int FILTER_NONE
      filter: No normalization/standardization
      See Also:
    • TAGS_FILTER

      public static final Tag[] TAGS_FILTER
      The filter to apply to the training data
  • Constructor Details

    • SMO

      public SMO()
  • Method Details

    • globalInfo

      public String globalInfo()
      Returns a string describing classifier
      Returns:
      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.
      Specified by:
      getTechnicalInformation in interface TechnicalInformationHandler
      Returns:
      the technical information about this class
    • turnChecksOff

      public void turnChecksOff()
      Turns off checks for missing values, etc. Use with caution.
    • turnChecksOn

      public void turnChecksOn()
      Turns on checks for missing values, etc.
    • getCapabilities

      public Capabilities getCapabilities()
      Returns default capabilities of the classifier.
      Specified by:
      getCapabilities in interface CapabilitiesHandler
      Specified by:
      getCapabilities in interface Classifier
      Overrides:
      getCapabilities in class AbstractClassifier
      Returns:
      the capabilities of this classifier
      See Also:
    • buildClassifier

      public void buildClassifier(Instances insts) throws Exception
      Method for building the classifier. Implements a one-against-one wrapper for multi-class problems.
      Specified by:
      buildClassifier in interface Classifier
      Parameters:
      insts - the set of training instances
      Throws:
      Exception - if the classifier can't be built successfully
    • distributionForInstance

      public double[] distributionForInstance(Instance inst) throws Exception
      Estimates class probabilities for given instance.
      Specified by:
      distributionForInstance in interface Classifier
      Overrides:
      distributionForInstance in class AbstractClassifier
      Parameters:
      inst - the instance to compute the probabilities for
      Returns:
      an array containing the estimated membership probabilities of the test instance in each class or the numeric prediction
      Throws:
      Exception - in case of an error
    • obtainVotes

      public int[] obtainVotes(Instance inst) throws Exception
      Returns an array of votes for the given instance.
      Parameters:
      inst - the instance
      Returns:
      array of votex
      Throws:
      Exception - if something goes wrong
    • sparseWeights

      public double[][][] sparseWeights()
      Returns the weights in sparse format.
    • sparseIndices

      public int[][][] sparseIndices()
      Returns the indices in sparse format.
    • bias

      public double[][] bias()
      Returns the bias of each binary SMO.
    • numClassAttributeValues

      public int numClassAttributeValues()
    • classAttributeNames

      public String[] classAttributeNames()
    • attributeNames

      public String[][][] attributeNames()
      Returns the attribute names.
    • listOptions

      public Enumeration<Option> listOptions()
      Returns an enumeration describing the available options.
      Specified by:
      listOptions in interface OptionHandler
      Overrides:
      listOptions in class AbstractClassifier
      Returns:
      an enumeration of all the available options.
    • setOptions

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

      Valid options are:

       -no-checks
          Turns off all checks - use with caution!
          Turning them off assumes that data is purely numeric, doesn't
          contain any missing values, and has a nominal class. Turning them
          off also means that no header information will be stored if the
          machine is linear. Finally, it also assumes that no instance has
          a weight equal to 0.
          (default: checks on)
       -C <double>
          The complexity constant C. (default 1)
       -N
          Whether to 0=normalize/1=standardize/2=neither. (default 0=normalize)
       -L <double>
          The tolerance parameter. (default 1.0e-3)
       -P <double>
          The epsilon for round-off error. (default 1.0e-12)
       -M
          Fit calibration models to SVM outputs. 
       -V <double>
          The number of folds for the internal
          cross-validation. (default -1, use training data)
       -W <double>
          The random number seed. (default 1)
       -K <classname and parameters>
          The Kernel to use.
          (default: weka.classifiers.functions.supportVector.PolyKernel)
       -calibrator <scheme specification>
          Full name of calibration model, followed by options.
          (default: "weka.classifiers.functions.Logistic")
       -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).
       
         Options specific to kernel weka.classifiers.functions.supportVector.PolyKernel:
         
       -E <num>
          The Exponent to use.
          (default: 1.0)
       -L
          Use lower-order terms.
          (default: no)
       -C <num>
          The size of the cache (a prime number), 0 for full cache and 
          -1 to turn it off.
          (default: 250007)
       -output-debug-info
          Enables debugging output (if available) to be printed.
          (default: off)
       -no-checks
          Turns off all checks - use with caution!
          (default: checks on)
       
         Options specific to calibrator weka.classifiers.functions.Logistic:
         
       -C
          Use conjugate gradient descent rather than BFGS updates.
       -R <ridge>
          Set the ridge in the log-likelihood.
       -M <number>
          Set the maximum number of iterations (default -1, until convergence).
       -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).
      Specified by:
      setOptions in interface OptionHandler
      Overrides:
      setOptions in class AbstractClassifier
      Parameters:
      options - the list of options as an array of strings
      Throws:
      Exception - if an option is not supported
    • getOptions

      public String[] getOptions()
      Gets the current settings of the classifier.
      Specified by:
      getOptions in interface OptionHandler
      Overrides:
      getOptions in class AbstractClassifier
      Returns:
      an array of strings suitable for passing to setOptions
    • setChecksTurnedOff

      public void setChecksTurnedOff(boolean value)
      Disables or enables the checks (which could be time-consuming). Use with caution!
      Parameters:
      value - if true turns off all checks
    • getChecksTurnedOff

      public boolean getChecksTurnedOff()
      Returns whether the checks are turned off or not.
      Returns:
      true if the checks are turned off
    • checksTurnedOffTipText

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

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

      public void setKernel(Kernel value)
      sets the kernel to use
      Parameters:
      value - the kernel to use
    • getKernel

      public Kernel getKernel()
      Returns the kernel to use
      Returns:
      the current kernel
    • calibratorTipText

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

      public void setCalibrator(Classifier value)
      sets the calibrator to use
      Parameters:
      value - the calibrator to use
    • getCalibrator

      public Classifier getCalibrator()
      Returns the calibrator to use
      Returns:
      the current calibrator
    • cTipText

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

      public double getC()
      Get the value of C.
      Returns:
      Value of C.
    • setC

      public void setC(double v)
      Set the value of C.
      Parameters:
      v - Value to assign to C.
    • toleranceParameterTipText

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

      public double getToleranceParameter()
      Get the value of tolerance parameter.
      Returns:
      Value of tolerance parameter.
    • setToleranceParameter

      public void setToleranceParameter(double v)
      Set the value of tolerance parameter.
      Parameters:
      v - Value to assign to tolerance parameter.
    • epsilonTipText

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

      public double getEpsilon()
      Get the value of epsilon.
      Returns:
      Value of epsilon.
    • setEpsilon

      public void setEpsilon(double v)
      Set the value of epsilon.
      Parameters:
      v - Value to assign to epsilon.
    • filterTypeTipText

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

      public SelectedTag getFilterType()
      Gets how the training data will be transformed. Will be one of FILTER_NORMALIZE, FILTER_STANDARDIZE, FILTER_NONE.
      Returns:
      the filtering mode
    • setFilterType

      public void setFilterType(SelectedTag newType)
      Sets how the training data will be transformed. Should be one of FILTER_NORMALIZE, FILTER_STANDARDIZE, FILTER_NONE.
      Parameters:
      newType - the new filtering mode
    • buildCalibrationModelsTipText

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

      public boolean getBuildCalibrationModels()
      Get the value of buildCalibrationModels.
      Returns:
      Value of buildCalibrationModels.
    • setBuildCalibrationModels

      public void setBuildCalibrationModels(boolean newbuildCalibrationModels)
      Set the value of buildCalibrationModels.
      Parameters:
      newbuildCalibrationModels - Value to assign to buildCalibrationModels.
    • numFoldsTipText

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

      public int getNumFolds()
      Get the value of numFolds.
      Returns:
      Value of numFolds.
    • setNumFolds

      public void setNumFolds(int newnumFolds)
      Set the value of numFolds.
      Parameters:
      newnumFolds - Value to assign to numFolds.
    • randomSeedTipText

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

      public int getRandomSeed()
      Get the value of randomSeed.
      Returns:
      Value of randomSeed.
    • setRandomSeed

      public void setRandomSeed(int newrandomSeed)
      Set the value of randomSeed.
      Parameters:
      newrandomSeed - Value to assign to randomSeed.
    • toString

      public String toString()
      Prints out the classifier.
      Overrides:
      toString in class Object
      Returns:
      a description of the classifier as a string
    • getRevision

      public String getRevision()
      Returns the revision string.
      Specified by:
      getRevision in interface RevisionHandler
      Overrides:
      getRevision in class AbstractClassifier
      Returns:
      the revision
    • main

      public static void main(String[] argv)
      Main method for testing this class.