Module es.upm.fi.cig.multictbnc
Package es.upm.fi.cig.multictbnc.conceptdriftdetection

Interface ConceptDriftScore

All Known Implementing Classes:
AverageLocalLogLikelihood
public interface ConceptDriftScore
Interface representing scores that can be computed locally on each node of a MultiCTBNC and used to detect concept drifts in a given data batch.

  • Method Summary

    Modifier and Type
    Method
    Description
    double[]
    compute(MultiCTBNC<CPTNode,CIMNode> model, Dataset newBatch, List<String> nameVariables)
    Compute the score over each node of the provided MultiCTBNC.

  • Method Details

    • compute

      double[] compute(MultiCTBNC<CPTNode,CIMNode> model, Dataset newBatch, List<String> nameVariables)
      Compute the score over each node of the provided MultiCTBNC.
      Parameters:
      model - MultiCTBNC model to be evaluated for concept drift
      newBatch - new data batch that is used to compute the drift score
      nameVariables - names of the variables
      Returns:
      array of doubles representing the drift scores for each node