Class OutlierSmROCCurve

  • All Implemented Interfaces:
    Evaluator, ResultProcessor

    @Reference(authors="W. Klement, P. A. Flach, N. Japkowicz, S. Matwin",
               title="Smooth Receiver Operating Characteristics (smROC) Curves",
               booktitle="European Conf. Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD\'11)",
               url="https://doi.org/10.1007/978-3-642-23783-6_13",
               bibkey="DBLP:conf/pkdd/KlementFJM11")
    public class OutlierSmROCCurve
    extends java.lang.Object
    implements Evaluator
    Smooth ROC curves are a variation of classic ROC curves that takes the scores into account.

    Reference:

    W. Klement, P. A. Flach, N. Japkowicz, S. Matwin
    Smooth Receiver Operating Characteristics (smROC) Curves.
    European Conf. Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD'11)

    However, this method has some deficiencies when the mean score is not 0.5, as discussed in:

    Erich Schubert, Remigius Wojdanowski, Arthur Zimek, Hans-Peter Kriegel
    On Evaluation of Outlier Rankings and Outlier Scores
    Proc. 12th SIAM Int. Conf. on Data Mining (SDM 2012)

    Since:
    0.5.0
    Author:
    Erich Schubert
    • Field Detail

      • SMAUROC_LABEL

        public static final java.lang.String SMAUROC_LABEL
        The label we use for marking AUROC values.
        See Also:
        Constant Field Values
      • LOG

        private static final Logging LOG
        The logger.
      • positiveClassName

        private java.util.regex.Pattern positiveClassName
        Stores the "positive" class.
    • Constructor Detail

      • OutlierSmROCCurve

        public OutlierSmROCCurve​(java.util.regex.Pattern positive_class_name)
        Constructor.
        Parameters:
        positive_class_name - Positive class name pattern