Aller au contenu Aller au menu Aller à la recherche

Événements et séminaires - LSTA
Laboratoire de Statistique Théorique et Appliquée


21/02/2017 - Anna Kiriliouk (UCL, Belgique)

 Groupe de travail théorie des valeurs extrêmes

Peaks over thresholds modelling with multivariate generalized Pareto distributions
(joint work with Holger Rootzén, Johan Segers and Jenny Wadsworth)

The multivariate generalized Pareto distribution arises as the limit of a suitably normalized vector conditioned upon at least one component of that vector being extreme. Statistical modelling using multivariate generalized Pareto distributions constitutes the multivariate analogue of peaks over thresholds modelling with the univariate generalized Pareto distribution. We introduce a construction device which allows us to develop a variety of new and existing parametric tail dependence models. A censored likelihood procedure is proposed to make inference on these models, together with several goodness-of-fit diagnostics. The models are fitted to rainfall data in the context of landslide risk estimation, showing that multivariate GP modelling can respect physical constraints in a way which was not possible before.