Dynamic Copula Mixtures with Application to the COVID-19 Pandemic

Details

Supervisors

Hafner, Christian

Faculty

Faculté des sciences

Degree label

Master [120] en statistique, orientation générale, à finalité spécialisée

Abstract

This thesis presents the development of a novel dynamic copula mixtures model designed to estimate time-varying dependencies between multiple time series. This model is a particular case of score-driven and dynamic copulas models. It was tested successfully on simulated data and was also applied to COVID-19 data from two neighbouring countries, focusing on growth rate of new cases. This application study revealed that, most of the time, there were independence or low dependence between the series. Despite some limitations related to the assumptions of normality and homoskedasticity, the model showed promising results when applied to real datasets. Future research could explore sensitivity analysis through simulation, test different copula functions and parameters, and apply the model to other datasets.