Welcome!This is where I jot down various things that I had to figure out on my own while using WinBUGS 1.4.1Impress your friendsI am a graphical interface/doodle user. I understand the WinBUGS code when I read it and can then make modifications to the code, but I am currently unable to generate code on my own. Since my models tend to have a lot of variables, I'll generate code structure from a doodle, then add in the full set of variables. I am not a statistician, I am merely a user (I work on health related quality of life measures). I have taken a full semester of graduate Bayesian coursework, so I am not flailing blindly, but I am not around a community of users to help me actually apply these methods to my problems.
The documentation for WinBUGS is notoriously poor. This is expecially true on the steep beginner learning curve. This site is not meant to be comprehensive, or even particularly organized, or a tutorial for getting started (it really is worth taking a crash course to get oriented). It is meant as a place to keep notes about quirks that have tripped me up. It is also meant as a place to store very generic models that actually run - I found the models included in the program to have so many levels when I first started out that I had a hard time figuring out what the core part of the model is.
I would be happy to post any tips or models other people would like to contribute.
If you're looking for more advanced discussion or help with a specific problem, it is always worth going through the BUGS archive.
The thing you need to solve all of your errors: a t-shirt (or mouse pad) with Bayes's Theorem on it! Available here!Common Error Messagesarray index is greater than array upper bound for [last variable in rectangle]WinBUGS Optimization and Data Storage
made use of undefined node
this chain contains uninitialized variables
incomplete row of data
multiple definitions of node [x]
incompatible copyData limits in WinBUGS?Other Notes
Simple OptimizationsUsing DIC for Model EvaluationGeneric Models
Variable names
Rectangular data structure
Batch scripts for WinBUGS
Out of sample testsLogistic Regression
Bernoulli Outcome
Nested Bernoulli as an alternative to ordinal logistic