Package opennlp.tools.ml.libsvm.doccat

Class SvmDoccatConfiguration.Builder

java.lang.Object
opennlp.tools.ml.libsvm.doccat.SvmDoccatConfiguration.Builder
Enclosing class:
SvmDoccatConfiguration
public static class SvmDoccatConfiguration.Builder extends Object
A builder for SvmDoccatConfiguration.

  • Constructor Details

    • Builder

      public Builder()

  • Method Details

    • setSvmConfiguration

      public SvmDoccatConfiguration.Builder setSvmConfiguration(de.hhn.mi.configuration.SvmConfiguration svmConfiguration)
      Sets the underlying SVM configuration.

      Note: Probability estimates must be enabled in the SVM configuration for the categorizer to produce probability distributions over categories.

      Parameters:
      svmConfiguration - The SVM configuration. Must not be null.
      Returns:
      This builder.

    • setTermWeightingStrategy

      public SvmDoccatConfiguration.Builder setTermWeightingStrategy(TermWeightingStrategy strategy)
      Sets the term weighting strategy.
      Parameters:
      strategy - The weighting strategy. Must not be null. Default: TermWeightingStrategy.TF_IDF.
      Returns:
      This builder.

    • setFeatureSelectionStrategy

      public SvmDoccatConfiguration.Builder setFeatureSelectionStrategy(FeatureSelectionStrategy strategy)
      Sets the feature selection strategy.
      Parameters:
      strategy - The selection strategy. Must not be null. Default: FeatureSelectionStrategy.NONE.
      Returns:
      This builder.

    • setMaxFeatures

      public SvmDoccatConfiguration.Builder setMaxFeatures(int maxFeatures)
      Sets the maximum number of features to retain after feature selection. Only applicable when FeatureSelectionStrategy is not FeatureSelectionStrategy.NONE.
      Parameters:
      maxFeatures - The maximum number of features. A value <= 0 means all features are retained. Default: -1.
      Returns:
      This builder.

    • setScaleFeatures

      public SvmDoccatConfiguration.Builder setScaleFeatures(boolean scaleFeatures)
      Sets whether feature values should be scaled to the range [lower, upper].
      Parameters:
      scaleFeatures - true to enable feature scaling. Default: true.
      Returns:
      This builder.

    • setScaleRange

      public SvmDoccatConfiguration.Builder setScaleRange(double lower, double upper)
      Sets the lower and upper bounds of the feature scaling range. Only applicable when setScaleFeatures(boolean) is true.
      Parameters:
      lower - The lower bound. Default: 0.0.
      upper - The upper bound. Must be greater than lower. Default: 1.0.
      Returns:
      This builder.
      Throws:
      IllegalArgumentException - if upper <= lower.

    • build

      public SvmDoccatConfiguration build()
      Returns:
      A fully configured SvmDoccatConfiguration.