Mémoires de Fin d’Etudes
Etablissement
Université de Biskra - Mohamed Khider
Affiliation
Département de Mathématique
Auteur
BETKA, Samah
Directeur de thèse
Djamel Meraghni (Maitre de conférence)
Filière
Mathématiques
Diplôme
Doctorat
Titre
Bivariate Extreme Value
Mots clés
Dependence; Bivariate extreme value index; Modeling tail; Asymptotic normality; Extreme quantile; Multivariate Extreme Value.
Résumé
The extreme value theory is the set of methods for statistical inference on the parameters of the generalized extreme value distribution and those of the generalized Pareto distribution as well as the extreme quantiles are estimated on the basis of this inference. In the one-dimensional case, there are a multitude of scientific work with applications in various fields such as insurance, finance, ... But in reality, the matters discussed are rather multidimensional in nature, hence the need to extend the inference techniques to the case of two dimensions and more. Modeling and analysing of the tail dependence is one of the numerous applications of this theory. There is a considerable amount of work within this framework. We are going to discuss, compare and improve the proposed estimators with emphasis on the bivariate case.
Réponse CS
Validé
Statut
Validé