Class PeaksOverThreshold

  • All Implemented Interfaces:
    MaximumLikelihoodFitting

    public class PeaksOverThreshold
    extends Object
    implements MaximumLikelihoodFitting
    Peaks Over Threshold (POT) method estimates the parameters for generalized Pareto distribution (GPD) using maximum likelihood on the observations that are over a given threshold.

    The R equivalent functions are evd::fpot.

    • Constructor Detail

      • PeaksOverThreshold

        public PeaksOverThreshold​(double threshold)
        Create an instance for POT method with a given threshold.

        The desired threshold is the lowest value for which the threshold excess fits the generalized Pareto distribution.

        Parameters:
        threshold - the threshold
    • Method Detail

      • fit

        public EstimateByLogLikelihood fit​(double[] observations)
                                    throws Exception
        Fits the observations to a generalized Pareto distribution (GPD).
        Specified by:
        fit in interface MaximumLikelihoodFitting
        Parameters:
        observations - the observations
        Returns:
        the estimate for the GPD parameters (scale, shape)
        Throws:
        Exception - if the optimization module failed to find parameters by MLE