Abstract
We describe a simple method for making inference on a functional of a multivariate distribution, based on its copula representation. We make use of an approximate Bayesian Monte Carlo algorithm, where the proposed values of the functional of interest are weighted in terms of their Bayesian exponentially tilted empirical likelihood. This method is particularly useful when the “true” likelihood function associated with the working model is too costly to evaluate or when the working model is only partially specified.
Highlights
Copula models are widely used in multivariate data analysis
A copula allows a useful representation of the joint distribution of a random vector in two steps: the marginal distributions and a distribution function which captures the dependence among the vector components
Bayesian Semi-Parametric Interpretation In Schennach (2005) it is argued and proved that the Bayesian exponentially tilted empirical likelihood has a precise Bayesian interpretation, which we describe in our context
Summary
Copula models are widely used in multivariate data analysis. Major areas of application include econometrics (Huynh et al, 2014), geophysics (Scholzel and Friederichs, 2008), climate prediction (Schefzik et al, 2013), actuarial science and finance (Cherubini et al, 2004), among the others. Its approximate posterior distribution is obtained via the use of the Bayesian exponentially tilted empirical likelihood approximation of the marginal likelihood of the quantity of interest, illustrated in Schennach (2005) This approximation of the true “unknown” likelihood function hopefully reduces the potential bias for incorrect distributional assumptions, very hard to check in complex dependence modeling.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
