Applied Bayesian Modelling, 2nd Edition

Preface

My gratitude is due to Wiley for proposing a revised edition of Applied Bayesian Modelling, first published in 2003. Much has changed since then for those seeking to apply Bayesian principles or to exploit the growing advantages of Bayesian estimation.

The central program used throughout the text in worked examples is BUGS, though R packages such as R-INLA, R2BayesX and MCMCpack are also demonstrated. Reference throughout the text to BUGS can be taken to refer both to WinBUGS and the ongoing OpenBUGS program, on which future development will concentrate (see http://www.openbugs.info/w/). There is a good deal of continuity between the final WinBUGS14 version and OpenBUGS (for details of differences see http://www.openbugs.info/w.cgi/OpenVsWin), though OpenBUGS has a wider range of sampling choices, distributions and functions. BUGS code can also be simply adapted to JAGS applications and the JAGS interfaces with R such as rjags.

Although R interfaces to BUGS or encapsulating the program are now widely used, the BUGS programming language itself remains a central aspect. Direct experience in WinBUGS or OpenBUGS programming is important as a preliminary to using R Interfaces such as BRUGS and rjags.

For learning Bayesian methods, especially if the main goal is data analysis per se, BUGS has advantages both practical and pedagogical. It can be seen as a half-way house between menu driven Bayesian computing (still not really established in any major computing package, though ...