Mixture Representation of the Maximum Entropy Density Through Archimedean Copulas

Abstract

In this article, we obtain a mixture representation of the maximum entropy density introduced by Rodrigues (2004) via Laplace approximation. This representation suggests, as in Sklar (1959), a dependence structure through Archimedean copulas independently of the specified marginal distributions. This result can be used as a natural Bayesian and non Bayesian procedure to estimate the dependence function and the marginal, separately.

Keywords:

• Archimedean copula
• Entropy model
• Kullback–Leibler distance
• Laplace approximation

Mathematics Subject Classification:

• Primary 94Axx
• Secondary 94A17

Acknowledgment

The first author was supported by CNPq-Grant: 300852/83-5(NV)