A0613
Title: Tail dependence with copulas
Authors: Fabrizio Durante - University of Salento (Italy) [presenting]
Abstract: Copula models have numerous advantages in describing the behavior of a multivariate stochastic system because of their flexibility in capturing various dependence aspects, especially in the tail of the joint distribution. We present several selected tools to describe the tail behaviour of random vectors by means of the copula approach. We start with a critical discussion about the classical tail dependence coefficients and their use. Then we consider some alternatives that are especially of interest in high dimensions. Computational aspects, as well as some applications, will be illustrated.
