Title: Multivariate regular variations and the Husler-Reiss Pareto model
Authors: Clement Dombry - Universite de Franche Comte (France) [presenting]
Olivier Zhen Wai Ho - Universite de Franche Comte (France)
Abstract: A simple procedure based on a multivariate Breiman Lemma is exposed for building multivariate regularly varying distributions. It encompasses the classical max-stable models such as logistic, negative logistic, Dirichlet, Husler-Reiss, extremal-Gaussian and extremal-t. In a second part, we focus on the Pareto model associated with the Husler-Reiss max-stable model. An interesting exponential family property is exhibited and maximum-likelihood estimation is considered. The case of different tail indices for different margins is also studied.