Title: Graphical modelling of extremes
Authors: Sebastian Engelke - Ecole Polytechnique Federale de Lausanne (Switzerland)
Robin Evans - University of Oxford (United Kingdom)
Adrien Hitz - University of Oxford (United Kingdom) [presenting]
Abstract: Graphical models provide a useful framework for inferring multivariate distributions in high dimensions by relying on conditional independence. We will explain how these ideas can be applied for modelling multivariate extreme observations. A valid limiting tail distribution must typically satisfy a homogeneity constraint, and we will see how Hammersley-Clifford theorem, a main result in graphical models, can be adapted to this context.