Bayesian Copula Regression

Tools for Bayesian copula generalized linear models (GLMs). The sampling scheme is based on Pitt, Chan, and Kohn (2006) . Regression parameters (including coefficients and dispersion parameters) are estimated via the adaptive random walk Metropolis approach developed by Haario, Saksman, and Tamminen (1999) . The prior for the correlation matrix is based on Hoff (2007) .

bayescopulareg

Bayesian analysis of multivariate generalized linear models via copulas

This package contains two main functions.

• bayescopulaglm samples from the posterior of a multivariate generalized linear model (currently supported families are poisson, gaussian, binomial, and gamma.
• predict.bayescopulaglm which samples from the predictive posterior density of a bayescopulaglm object.