Class GISTrainer

java.lang.Object
opennlp.tools.ml.AbstractTrainer<opennlp.tools.util.TrainingParameters>
opennlp.tools.ml.AbstractEventTrainer<opennlp.tools.util.TrainingParameters>
opennlp.tools.ml.maxent.GISTrainer
All Implemented Interfaces:
opennlp.tools.commons.Trainer<opennlp.tools.util.TrainingParameters>, opennlp.tools.ml.EventTrainer<opennlp.tools.util.TrainingParameters>
public class GISTrainer extends opennlp.tools.ml.AbstractEventTrainer<opennlp.tools.util.TrainingParameters>
An implementation of Generalized Iterative Scaling (GIS).

The reference paper for this implementation was Adwait Ratnaparkhi's tech report at the University of Pennsylvania's Institute for Research in Cognitive Science, and is available at ftp://ftp.cis.upenn.edu/pub/ircs/tr/97-08.ps.Z.

The slack parameter used in the above implementation has been removed by default from the computation and a method for updating with Gaussian smoothing has been added per Investigating GIS and Smoothing for Maximum Entropy Taggers, Clark and Curran (2002). http://acl.ldc.upenn.edu/E/E03/E03-1071.pdf.

The slack parameter can be used by setting useSlackParameter to true. Gaussian smoothing can be used by setting useGaussianSmoothing to true.

A Prior can be used to train models which converge to the distribution which minimizes the relative entropy between the distribution specified by the empirical constraints of the training data and the specified prior. By default, the uniform distribution is used as the prior.

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    static final double
    LOG_LIKELIHOOD_THRESHOLD_DEFAULT
     
    static final String
    LOG_LIKELIHOOD_THRESHOLD_PARAM
     

    Fields inherited from class opennlp.tools.ml.AbstractEventTrainer

    DATA_INDEXER_ONE_PASS_REAL_VALUE, DATA_INDEXER_ONE_PASS_VALUE, DATA_INDEXER_PARAM, DATA_INDEXER_TWO_PASS_VALUE

    Fields inherited from interface opennlp.tools.ml.EventTrainer

    EVENT_VALUE

  • Constructor Summary

    Constructors
    Constructor
    Description
    GISTrainer()
    Initializes a GISTrainer.

  • Method Summary

    Modifier and Type
    Method
    Description
    opennlp.tools.ml.model.MaxentModel
    doTrain(opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> indexer)
    void
    init(opennlp.tools.util.TrainingParameters trainingParameters, Map<String,String> reportMap)
    boolean
    isSortAndMerge()
    void
    setGaussianSigma(double sigmaValue)
    Sets whether this trainer will use smoothing while training the model.
    void
    setSmoothing(boolean smooth)
    Sets whether this trainer will use smoothing while training the model.
    void
    setSmoothingObservation(double timesSeen)
    Sets whether this trainer will use smoothing while training the model.
    GISModel
    trainModel(int iterations, opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> di)
    Trains a model using the GIS algorithm.
    GISModel
    trainModel(int iterations, opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> di, int threads)
    Trains a model using the GIS algorithm.
    GISModel
    trainModel(int iterations, opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> di, opennlp.tools.ml.model.Prior modelPrior, int threads)
    Trains a model using the GIS algorithm.
    GISModel
    trainModel(opennlp.tools.util.ObjectStream<opennlp.tools.ml.model.Event> eventStream)
    Trains a model using the GIS algorithm, assuming 100 iterations and no cutoff.
    GISModel
    trainModel(opennlp.tools.util.ObjectStream<opennlp.tools.ml.model.Event> eventStream, int iterations, int cutoff)
    Trains a GIS model on the event in the specified event stream, using the specified number of iterations and the specified count cutoff.

    Methods inherited from class opennlp.tools.ml.AbstractEventTrainer

    getDataIndexer, train, train, validate

    Methods inherited from class opennlp.tools.ml.AbstractTrainer

    getAlgorithm, getCutoff, getIterations, getTrainingConfiguration, init

    Methods inherited from class Object

    equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    Methods inherited from interface opennlp.tools.commons.Trainer

    init

  • Field Details

    • LOG_LIKELIHOOD_THRESHOLD_PARAM

      public static final String LOG_LIKELIHOOD_THRESHOLD_PARAM
      See Also:

    • LOG_LIKELIHOOD_THRESHOLD_DEFAULT

      public static final double LOG_LIKELIHOOD_THRESHOLD_DEFAULT
      See Also:

  • Constructor Details

    • GISTrainer

      public GISTrainer()
      Initializes a GISTrainer.

      Note:
      The resulting instance does not print progress messages about training to STDOUT.

  • Method Details

    • isSortAndMerge

      public boolean isSortAndMerge()
      Specified by:
      isSortAndMerge in class opennlp.tools.ml.AbstractEventTrainer<opennlp.tools.util.TrainingParameters>

    • init

      public void init(opennlp.tools.util.TrainingParameters trainingParameters, Map<String,String> reportMap)
      Specified by:
      init in interface opennlp.tools.commons.Trainer<opennlp.tools.util.TrainingParameters>
      Overrides:
      init in class opennlp.tools.ml.AbstractTrainer<opennlp.tools.util.TrainingParameters>

    • doTrain

      public opennlp.tools.ml.model.MaxentModel doTrain(opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> indexer) throws IOException
      Specified by:
      doTrain in class opennlp.tools.ml.AbstractEventTrainer<opennlp.tools.util.TrainingParameters>
      Throws:
      IOException

    • setSmoothing

      public void setSmoothing(boolean smooth)
      Sets whether this trainer will use smoothing while training the model.

      Note:
      This can improve model accuracy, though training will potentially take longer and use more memory. Model size will also be larger.

      Parameters:
      smooth - true if smoothing is desired, false if not.

    • setSmoothingObservation

      public void setSmoothingObservation(double timesSeen)
      Sets whether this trainer will use smoothing while training the model.

      Note:
      This can improve model accuracy, though training will potentially take longer and use more memory. Model size will also be larger.

      Parameters:
      timesSeen - The "number" of times we want the trainer to imagine it saw a feature that it actually didn't see

    • setGaussianSigma

      public void setGaussianSigma(double sigmaValue)
      Sets whether this trainer will use smoothing while training the model.

      Note:
      This can improve model accuracy, though training will potentially take longer and use more memory. Model size will also be larger.

      Parameters:
      sigmaValue - The Gaussian sigma value used for smoothing.

    • trainModel

      public GISModel trainModel(opennlp.tools.util.ObjectStream<opennlp.tools.ml.model.Event> eventStream) throws IOException
      Trains a model using the GIS algorithm, assuming 100 iterations and no cutoff.
      Parameters:
      eventStream - The eventStream holding the data on which this model will be trained.
      Returns:
      A trained GISModel which can be used immediately or saved to disk using an GISModelWriter.
      Throws:
      IOException

    • trainModel

      public GISModel trainModel(opennlp.tools.util.ObjectStream<opennlp.tools.ml.model.Event> eventStream, int iterations, int cutoff) throws IOException
      Trains a GIS model on the event in the specified event stream, using the specified number of iterations and the specified count cutoff.
      Parameters:
      eventStream - A stream of all events.
      iterations - The number of iterations to use for GIS.
      cutoff - The number of times a feature must occur to be included.
      Returns:
      A trained GISModel which can be used immediately or saved to disk using an GISModelWriter.
      Throws:
      IOException

    • trainModel

      public GISModel trainModel(int iterations, opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> di)
      Trains a model using the GIS algorithm.
      Parameters:
      iterations - The number of GIS iterations to perform.
      di - The DataIndexer used to compress events in memory.
      Returns:
      A trained GISModel which can be used immediately or saved to disk using an GISModelWriter.
      Throws:
      IllegalArgumentException - Thrown if parameters were invalid.

    • trainModel

      public GISModel trainModel(int iterations, opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> di, int threads)
      Trains a model using the GIS algorithm.
      Parameters:
      iterations - The number of GIS iterations to perform.
      di - The DataIndexer used to compress events in memory.
      threads - The number of thread to train with. Must be greater than 0.
      Returns:
      A trained GISModel which can be used immediately or saved to disk using an GISModelWriter.
      Throws:
      IllegalArgumentException - Thrown if parameters were invalid.

    • trainModel

      public GISModel trainModel(int iterations, opennlp.tools.ml.model.DataIndexer<opennlp.tools.util.TrainingParameters> di, opennlp.tools.ml.model.Prior modelPrior, int threads)
      Trains a model using the GIS algorithm.
      Parameters:
      iterations - The number of GIS iterations to perform.
      di - The DataIndexer used to compress events in memory.
      modelPrior - The Prior distribution used to train this model.
      Returns:
      A trained GISModel which can be used immediately or saved to disk using an GISModelWriter.
      Throws:
      IllegalArgumentException - Thrown if parameters were invalid.