A semiparametric family of bivariate copulas: dependence properties and estimation procedures
Résumé
Copulas are a relevant tool to build multivariate probability laws, from fixed margins and required degree of dependence. In this communication, we propose simple estimation methods dedicated to a semiparametric family of bivariate copulas. These copulas can be simply estimated through the estimation of their univariate generating function. We take profit of this result to estimate the associated measures of association as well as the high probability regions of the copula. These procedures are illustrated on simulations and on real data.
Domaines
Méthodologie [stat.ME]
Origine : Fichiers produits par l'(les) auteur(s)