Class NumberCriterion

  • All Implemented Interfaces:
    IMeasure, IPriorMeasure

    public class NumberCriterion
    extends AbstractSignificanceBasedMeasure
    implements IPriorMeasure
    The N/Number/AvgC criterion is a prior efficiency measure based on the size of the label set. See [V. Vovk, V. Fedorova, I. Nouretdinov and A. Gammerman, "Criteria of Efficiency for Conformal Prediction", COPA 2016, LNAI 9653, pp. 23-39, 2016] for the definitions used here.
    Author:
    anders.gidenstam(at)hb.se
    • Constructor Detail

      • NumberCriterion

        public NumberCriterion​(double significanceLevel)
        Creates a N/Number criterion measure.
        Parameters:
        significanceLevel - the significance level used for the label sets.
    • Method Detail

      • compute

        public double compute​(ConformalClassification prediction)
        Computes the N/Number criterion measure for this prediction.
        Specified by:
        compute in interface IPriorMeasure
        Parameters:
        prediction - the prediction.
        Returns:
        the Number criterion measure for this prediction. Small values are preferable.