Package ai.djl.basicdataset.tabular.utils

Interface Featurizer

All Known Subinterfaces:
Featurizer.DataFeaturizer, PreparedFeaturizer
public interface Featurizer
An interface that convert String to numeric data.

  • Nested Class Summary

    Nested Classes
    Modifier and Type
    Interface
    Description
    static interface 
    Featurizer.DataFeaturizer
    A Featurizer that only supports the data featurize operations, but not the full deFeaturize operations used by labels.

  • Method Summary

    Modifier and Type
    Method
    Description
    int
    dataRequired()
    Returns the length of the data array required by deFeaturize(float[]).
    Object
    deFeaturize(float[] data)
    Converts the output data for a label back into the Java type.
    void
    featurize(DynamicBuffer buf, String input)
    Puts encoded data into the float buffer.

  • Method Details

    • featurize

      void featurize(DynamicBuffer buf, String input)
      Puts encoded data into the float buffer.
      Parameters:
      buf - the float buffer to be filled
      input - the string input

    • dataRequired

      int dataRequired()
      Returns the length of the data array required by deFeaturize(float[]).
      Returns:
      the length of the data array required by deFeaturize(float[])

    • deFeaturize

      Object deFeaturize(float[] data)
      Converts the output data for a label back into the Java type.
      Parameters:
      data - the data vector correspondign to the feature
      Returns:
      a Java type (depending on the Featurizer) representing the data.