Bassist — a tool for MCMC simulation of statistical models

Bassist is a simulation tool for the analysis of complex statistical models. Given a high-level specification of a full probability model, Bassist generates a simulator for analysing the model with respect to given data. Bassist follows the Bayesian modeling approach. Model parameters and other unknown quantities are assigned prior distributions by the modeler; the likelihood of observed data follows from the specification of the statistical model. The matter of interest is the posterior probability of the model, i.e., the joint probability distribution of all unknown quantities conditioned on the observed data. Bayesian statistics provide a wellfounded means for assigning probabilities for models. For most realistic models, however, analytical methods for the computation of the posterior distribution are not feasible. To solve the problem, Bassist applies Markov chain Monte Carlo (MCMC) simulation techniques to approximate a sample from the joint posterior distribution.