Multivariate Peaks-over-Threshold Modelling for Spatial Extreme
Events
Description
Tools for high-dimensional peaks-over-threshold inference and simulation of Brown-Resnick and extremal Student spatial extremal processes. These include optimization routines based on censored likelihood and gradient scoring, and exact simulation algorithms for max-stable and multivariate Pareto distributions based on rejection sampling. Fast multivariate Gaussian and Student distribution functions using separation-of-variable algorithm with quasi Monte Carlo integration are also provided. Key references include de Fondeville and Davison (2018) , Thibaud and Opitz (2015) , Wadsworth and Tawn (2014) and Genz and Bretz (2009) .