Analysis of multidimensional probability distributions with copula functions. II

This article contains the second part of the consultation series on copula functions and their use in modeling multidimensional probability distributions. It describes pair-copula functions (including the concept of canonical and D-vines), alternative measures of dependence useful to summarize the dependence structure of the analyzed variables (including measures of tail dependence, particularly relevant in the case of asymmetric distributions), as well as parametric, semi-parametric and nonparametric methods of statistical estimation of copula functions.