Bayesian inference for multivariate extreme value distributions (Q1684166)

Max-stable distributions and processes are used for statistical modeling of multivariate and spatial extreme events. \textit{E. Thibaud} and \textit{T. Opitz} [Biometrika 102, No. 4, 855--870 (2015; Zbl 1372.62011)] used a Bayesian approach to fit a Brown-Resnick process to extreme temperatures. The authors extend this idea to develop methods applicable to general max-stable distributions that use full likelihoods. They prove a Bernstein-von Mises theorem under some conditions and prove the asymptotic normality of the median of the posterior distribution under some conditions for some multivariate and spatial extreme value statistics.