A NEW CLASS OF COPULAS WITH TAIL DEPENDENCE

摘要

Copula-based multivariate models allow to specify the marginal distributions separately from the dependence structure (i.e. the copula) which links these distributions to form a joint distribution. Within this work we introduce a new family of copulas ("generalized mean copulas") which is positive comprehensive (i.e. includes both independence and maximum dependence) and allows for upper tail dependence. It includes the Spearman copula and a specific Frechet copula as special cases. Some properties of the new copula are derived.