Comparative study of copula families via simulations


This paper provides an introduction to copulas models, specifically the empirical  and semi-empirical copulas, dedicated to evaluate the advantages and disadvantages of adopting empirical copulas, based on nonparametric marginal distributions of the involved variables, and the semiempirical which resort to parametric adjustments for the marginals distributions. Simulation data from linear and non-linear dependent variables, as well as independent variables that follow exponential distributions, illustrate the similarities between these alternatives of copulas models. It is evident from the study, the good results of the empirical copula in relation to the semi-empirical, as well as the need for more in-depth studies on the contributions of copulas models.

Probability and Statistics