Change-point analysis of copula models
schedule le lundi 06 janvier 2020 de 17h00 à 18h00
Organisé par : F. Bechtold, W. Da Silva , A. Fermanian, S. Has, Y. Yu
Intervenant : Karen A. Vásquez Vivas (LPSM)
Lieu : Jussieu 16-26-209
Sujet : Change-point analysis of copula models
Copula analysis has become a widely used tool
to model the dependence structure of a vector of continuous random
variables. The success of this approach finds its origin in Sklar's
Theorem, which shows that the joint distribution function can be
expressed in terms of the marginal distribution functions and a copula.
In this talk I will present a copula model for variables that evolve through time. The copula parameter will be subject to abrupt changes. Neither the location nor the size of these changes are assumed to be known. The estimation is performed through the maximization of a likelihood criterion. Then I will propose a model selection procedure to determine the appropriate number of changes that allow to describe correctly the data after a simulation study with different settings. Finally, I will illustrate the estimation and model selection procedures applying them to real data.