Thesis title: Analisi della dipendenza tramite copule in presenza di dati misti: applicazioni in ambito attuariale
Since Sklar's work, copulas have increasingly become a fundamental tool for studying and modelling dependence structures between random variables, finding increasing application in various contexts, including actuarial.
This work has a twofold objective. From an applied perspective, it aims to expand the actuarial literature on the study of Policyholder Behaviour (PHB) by providing two applications on real data that demonstrate the potential of the copula-based approach with respect to market practices, particularly in the study of the phenomenon of redemptions.
From a theoretical perspective, the goal is to extend the traditional framework, enabling the use of copulas in the modelling of nominal categorical variables. A critical comparative analysis of the most commonly used methodologies for managing these variables is provided, and a solution is proposed for their consistent and meaningful inclusion.