Multivariate peaks over thresholds models

TL;DR

This paper advances multivariate peaks over thresholds modeling by developing theoretical foundations, flexible representations, inference tools, and simulation algorithms to enable higher-dimensional applications of generalized Pareto distributions.

Contribution

It introduces a general point process model for extreme episodes and provides multiple representations, inference methods, and simulation techniques for multivariate generalized Pareto distributions.

Findings

Derived a comprehensive point process model for extreme episodes.

Developed numerically tractable density and censored density forms.

Created new simulation algorithms and explicit probability formulas.

Abstract

Multivariate peaks over thresholds modeling based on generalized Pareto distributions has up to now only been used in few and mostly 2-dimensional situations. This paper contributes theoretical understanding, physically based models, inference tools, and simulation methods to support routine use, with an aim at higher dimensions. We derive a general point process model for extreme episodes in data, and show how conditioning the distribution of extreme episodes on threshold exceedance gives four basic representations of the family of generalized Pareto distributions. The first representation is constructed on the real scale of the observations. The second one starts with a model on a standard exponential scale which then is transformed to the real scale. The third and fourth are reformulations of a spectral representation proposed in A. Ferreira and L. de Haan [Bernoulli 20 (2014) 1717--1737]. Numerically tractable forms of densities and censored densities are found and give tools for flexible parametric likelihood inference. New simulation algorithms, explicit formulas for probabilities and conditional probabilities, and conditions which make the conditional distribution of weighted component sums generalized Pareto are derived.