Calendrier des événements

Partager sur :

WG Risk : Geometric approaches to statistics of multivariate extremes, with application to river flows

The Working Group on Risk - CREAR, with the support of the ESSEC IDS dpt, Institut des Actuaires, Fondation des Sciences de la Modélisation (CY - Labex MME-DII), and the group Risques AEF - SFdS, has the pleasure to invite you to the seminar by :

 

Prof. Jennifer Wadsworth

 

Dr. Jennifer Wadsworth is a professor of statistics at Lancaster University, UK. She completed her PhD at Lancaster in 2012, followed by postdoc positions at EPFL, Switzerland, and the University of Cambridge, UK. Her research focuses on extreme value theory, particularly multivariate and spatial modelling, with environmental applications. She currently supervises four PhD students and two postdocs, and a current particular focus of the group is a new geometric perspective on extremal dependence.

 

“Geometric approaches to statistics of multivariate extremes, with application to river flows”

 A geometric representation for multivariate extremes, based on the shapes of scaled sample clouds in light-tailed margins and their so-called limit sets, has recently been shown to connect several existing extremal dependence concepts. However, these results are purely probabilistic, and the geometric approach itself has not been fully exploited for statistical inference. We outline a method for parametric estimation of the limit set shape, which includes a useful non-/semi-parametric estimate as a pre-processing step. More fundamentally, our approach provides a new class of asymptotically motivated statistical models for the tails of multivariate distributions, and such models can accommodate any combination of simultaneous or non-simultaneous extremes through appropriate parametric forms for the limit set shape. In this talk we will also present ongoing work moving towards semiparametric methodology for greater flexibility. Extrapolation further into the tail of the distribution is possible via simulation from the fitted model, and probability estimates are possible in regions where other frameworks struggle. The methods are applied to river flow measurements in north-west England.

 

Dual format:

ESSEC Paris La Défense (CNIT), Room to confirm, and via Zoomplease click here .

 

Mercredi 30 octobre 2024
12h30 - 18h30 (GMT +2)
L'événement est organisé en présentiel et en ligne
ESSEC Paris La Défense
2 Pl. de la Défense
92800 Puteaux
En ligne
Lieu

ESSEC Paris La Défense

2 Pl. de la Défense
92800 Puteaux

Aucun commentaire

Vous devez être connecté pour laisser un commentaire. Connectez-vous.

Mercredi 30 octobre 2024
12h30 - 18h30 (GMT +2)
L'événement est organisé en présentiel et en ligne
ESSEC Paris La Défense
2 Pl. de la Défense
92800 Puteaux
En ligne
  • Ajouter à mon agenda