CMStatistics 2018: Start Registration
View Submission - CMStatistics
B0291
Title: Parametric assumptions in extreme value theory Authors:  Jenny Wadsworth - Lancaster University (United Kingdom) [presenting]
Abstract: In univariate extreme value theory, parametric limiting distributions for the tail arise under weak conditions on the underlying distribution. There is no such luck for the multivariate or spatial case: although non-degenerate limit distributions often arise, and certain conditions must be satisfied, there is no finite-dimensional parameterization of all possible dependence structures. However, parametric assumptions are a common simplification and often necessary for tackling high-dimensional problems. We will review some of the different approaches to multivariate extremes and aim to discuss the relative merits of (non) parametric inference methods.