Interface ClusteringEvaluation

All Superinterfaces:
Evaluation<ClusterID>, com.oracle.labs.mlrg.olcut.provenance.Provenancable<EvaluationProvenance>
public interface ClusteringEvaluation extends Evaluation<ClusterID>
An Evaluation for clustering tasks.

  • Method Summary

    Modifier and Type
    Method
    Description
    double
    adjustedMI()
    Measures the adjusted normalized mutual information between the predicted ids and the supplied ids.
    double
    normalizedMI()
    Calculates the normalized MI between the ground truth clustering ids and the predicted ones.

    Methods inherited from interface org.tribuo.evaluation.Evaluation

    asMap, get, getPredictions

    Methods inherited from interface com.oracle.labs.mlrg.olcut.provenance.Provenancable

    getProvenance

  • Method Details

    • normalizedMI

      double normalizedMI()
      Calculates the normalized MI between the ground truth clustering ids and the predicted ones.

      The value is bounded between 0 and 1.

      If this value is 1, then the predicted id values are a permutation of the supplied ids. If the value is 0 then the predicted ids are random wrt the supplied ids.

      Returns:
      The normalized MI.

    • adjustedMI

      double adjustedMI()
      Measures the adjusted normalized mutual information between the predicted ids and the supplied ids.

      The value is bounded between 0 and 1.

      If this value is 1, then the predicted id values are a permutation of the supplied ids. If the value is 0 then the predicted ids are random wrt the supplied ids.

      It's adjusted for chance unlike the normalized one.

      Returns:
      The adjusted MI.