bang

Bayesian Analysis, No Gibbs

What does bang do?

Provides functions for the Bayesian analysis of some simple commonly-used models, without using Markov Chain Monte Carlo (MCMC) methods such as Gibbs sampling. The ‘rust’ package https://cran.r-project.org/package=rust is used to simulate a random sample from the required posterior distribution, using the ratio-of-uniforms method. Currently three conjugate hierarchical models are available: beta-binomial, gamma-Poisson and a 1-way Analysis of Variance (ANOVA). Advantages of the ratio-of-uniforms method over MCMC in this context are that the user is not required to set tuning parameters nor to monitor convergence and a random posterior sample is produced.

A simple example

The hef function samples from the posterior distribution of the parameters of certain hierarchical exponential family models. The following code performs essentially the same analysis of the rat tumor data using a beta-binomial hierarchical model that appears in Section 5.3 of Gelman, A., Carlin, J. B., Stern, H. S. Dunson, D. B., Vehtari, A. and Rubin, D. B. (2014) Bayesian Data Analysis. Chapman & Hall / CRC. http://www.stat.columbia.edu/~gelman/book/.

library(bang)
rat_res <- hef(model = "beta_binom", data = rat)
plot(rat_res)

Installation

To get the current released version from CRAN:

install.packages("bang")