Lundi soir, Arthur Charpentier donnera un exposé au Groupe de Travail Big Data de l’Institut des Actuaires. Les slides e l’exposé sont en ligne,
Archives de catégorie : Talks
Spatial Aspects of Genealogical Data
This week, Ewen will be in Portugal at GEOMED 2017, the international, interdisciplinary conference on spatial statistics, geographical epidemiology and geographical aspects of public health, to present preliminary results on demographic datasets he has been working on this Summer.
Pooling Risks in a Community
Cette semaine, Arnaud Goussebaile présentera ses travaux à la conférence de l’association for public economic theory, sur pooling natural catastophe risks in a community (les slides sont en ligne)
Sinon Arnaud a obtenu le Ernst Meyer Prize la semaine passée… félicitations !
Matinée de la chaire, code collaboratif
Mardi matin, Christophe Dutang est venu faire une présentation sur le code collaboratif (à partir de son expérience personnelle)
Striving now to shirk later? intertemporal effort in car insurance
On Tuesday 31st of January, Alexandre Godzinski présenter will give a talk on adverse selection in car insurance, “striving now to shirk later? intertemporal effort in car insurance“. As always, there will also be someone from Covéa presenting recent actuarial results, with a practioner’s perspective.
Talk on Telematics
This morning, Katrien Antonio gave a talk on Telematic Data. The slides will be uploaded soon. The talk was based on the working paper Unraveling the predictive power of telematics data in car insurance pricing with R. Verbelen and G. Claeskens.
2nd Pricing Game à 100% Data Science
Cet après midi, Arthur Charpentier présentait à 100% les conclusions du second pricing game.
Accidents Corporels: Tribunal ou Assureur?
Ce mardi, Arthur Charpentier présentera un travail mené avec E. Blez, P.Y. Geoffard et J. Tomas sur la modélisation de la décision d’aller au tribunal lors d’un accident corporel. Il s’agit de résultats préliminaires. Le papier et les slides seront mis en ligne dans les semaines à venir.
Tree-based censored regression
Tuesday morning, December 8th, first talk related to the ACTINFO Chair, at Covéa, by Xavier Milhaud.
In this paper, we propose a regression tree procedure to estimate the conditional distribution of a variable which is not directly observed due to censoring. The model that we consider is motivated by applications in insurance, including the analysis of guarantees that involve durations, and claim reserving. We derive consistency results for our procedure, and for the selection of an optimal subtree using a pruning strategy. These theoretical results are supported by a simulation study, and two applications to insurance datasets. The first one concerns income protection insurance, while the second deals with reserving in third-party liability insurance.
Paper is available on https://hal.archives-ouvertes.fr.