Class MultiClassificationPerformance

java.lang.Object
adaa.analytics.rules.logic.performance.AbstractPerformanceCounter
adaa.analytics.rules.logic.performance.MultiClassificationPerformance

public class MultiClassificationPerformance extends AbstractPerformanceCounter
Measures the accuracy and classification error for both binary classification problems and multi class problems. Additionally, this performance criterion can also compute the kappa statistics for multi class problems. This is calculated as k = (P(A) - P(E)) / (1 - P(E)) with [ P(A) = diagonal sum / number of examples ] and [ P(E) = sum over i of ((sum of i-th row * sum of i-th column) / (n to the power of 2) ].
  • Field Details

    • ACCURACY

      public static final int ACCURACY
      Indicates accuracy.
      See Also:
    • ERROR

      public static final int ERROR
      Indicates classification error.
      See Also:
    • KAPPA

      public static final int KAPPA
      Indicates kappa statistics.
      See Also:
    • NAMES

      public static final String[] NAMES
      The names of the criteria.
  • Constructor Details

    • MultiClassificationPerformance

      public MultiClassificationPerformance(int type)
      Creates a MultiClassificationPerformance with the given type.
  • Method Details