Package ai.libs.jaicore.ml.core.evaluation.measure.multilabel

Class AutoMEKAGGPFitnessMeasureLoss

  • All Implemented Interfaces:
    IMeasure<double[],​java.lang.Double>, IMultilabelMeasure
    public class AutoMEKAGGPFitnessMeasureLoss
extends ADecomposableMultilabelMeasure
    Measure combining exact match, hamming loss, f1macroavgL and rankloss. Here implemented in inverse. de Sá, Alex GC, Gisele L. Pappa, and Alex A. Freitas. "Towards a method for automatically selecting and configuring multi-label classification algorithms." Proceedings of the Genetic and Evolutionary Computation Conference Companion. ACM, 2017.

    • Method Summary

      All Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      java.lang.Double calculateAvgMeasure​(java.util.List<double[]> actual, java.util.List<double[]> expected)
      Computes the measure for lists of input actual and the expected outcome expected and aggregates the measured values with the mean, as this is the most frequently used aggregate function.
      java.lang.Double calculateMeasure​(double[] actual, double[] expected)
      Computes the measure for a measured input actual and the expected outcome expected.

      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    • Constructor Detail

      • AutoMEKAGGPFitnessMeasureLoss

        public AutoMEKAGGPFitnessMeasureLoss()

    • Method Detail

      • calculateMeasure

        public java.lang.Double calculateMeasure​(double[] actual,
                                         double[] expected)
        Description copied from interface: IMeasure
        Computes the measure for a measured input actual and the expected outcome expected.
        Parameters:
        actual - The actually available values.
        expected - The expected values to compare the actual values with.
        Returns:
        The return value of the measure.

      • calculateAvgMeasure

        public java.lang.Double calculateAvgMeasure​(java.util.List<double[]> actual,
                                            java.util.List<double[]> expected)
        Description copied from interface: IMeasure
        Computes the measure for lists of input actual and the expected outcome expected and aggregates the measured values with the mean, as this is the most frequently used aggregate function.
        Specified by:
        calculateAvgMeasure in interface IMeasure<double[],​java.lang.Double>
        Overrides:
        calculateAvgMeasure in class ADecomposableDoubleMeasure<double[]>
        Parameters:
        actual - The list of actually available values.
        expected - The list of expected values to compare the actual values with.
        Returns:
        The mean of return values as output by the instance-wise measure.