A supermartingale argument for characterizing the Functional Hill process weak law for small parameters

TL;DR

This paper uses supermartingale techniques to characterize the weak limit laws of the Functional Hill process for small parameters, focusing on extreme value index estimators within Weibull domains.

Contribution

It introduces a supermartingale-based approach to derive the non-Gaussian asymptotic distribution of the Functional Hill process for small parameters, extending existing methods.

Findings

Supermartingale methods effectively characterize the asymptotic distribution.

Simulation results support the theoretical findings.

Statistical tests demonstrate practical applicability.

Abstract

The paper deals with the asymptotic laws of functional of standard random variables. These classes of statistics are closely related to estimators of the extreme value index when the underlying distribution function is in the Weibull domain of attraction. We use techniques based on martingales theory to describe the non Gaussian asymptotic distribution of the aforementioned statistics. We provide results of a simulation study as well as statistical tests that may be of interest with the proposed results.