An approach to analyze the tail of the distribution of heterogenous data and application to insurance

schedule le lundi 11 mai 2020 de 17h00 à 18h00

Organisé par : F. Bechtold, W. Da Silva , A. Fermanian, S. Has, Y. Yu

Intervenant : Sébastien Farkas (LPSM)
Lieu : online at https://bigbluebutton.math.upmc.fr/b/ade-phf-9dg

Sujet : An approach to analyze the tail of the distribution of heterogenous data and application to insurance

Résumé :

We propose a method based on regression trees to analyze the tail of the distribution of heterogenous data. Focusing on observations that are above a given threshold that we called severe/extreme data, we  combine a Generalized Pareto modeling - legitimate from Extreme Value Theory - and the regression tree (CART) approach. We will comment on the performence of the methodology based on simulations and then illustrate it on a public database about data breaches which can be helpful for insurers to quantify risks that emerged with the rise of the digital economy, called cyber risks.