ABSTRACT
This chapter provides a general perspective on computation for Bayesian data analysis. Our computational strategy for complex problems begins with crude initial estimates of model parameters and usually continues with Markov chain simulation, sometimes using mode-based approximations of the posterior distribution as an intermediate step. We begin with some notation and terminology that are relevant to all the chapters of this part of the book. Chapters 11, 12, and 13 cover Markov chain simulation, mode-based approximations, and more advanced methods in Bayesian computation.
