Variable Selection in a Hierarchical Bayesian Model for a Hazard Function
Fits a piecewise exponential hazard to survival data using a Hierarchical Bayesian model with an Intrinsic Conditional Autoregressive formulation for the spatial dependency in the hazard rates for each piece. This function uses Metropolis- Hastings-Green MCMC to allow the number of split points to vary and also uses Stochastic Search Variable Selection to determine what covariates drive the risk of the event. This function outputs trace plots depicting the number of split points in the hazard and the number of variables included in the hazard. The function saves all posterior quantities to the desired path.