Run OpenBUGS from R
The bugs function takes data and starting values as input. It automatically writes a OpenBUGS script, calls the model, and saves the simulations for easy access in .
bugs(data, inits, parameters.to.save, n.iter, model.file="model.txt", n.chains=3, n.burnin=floor(n.iter / 2), n.thin=1, saveExec=FALSE,restart=FALSE, debug=FALSE, DIC=TRUE, digits=5, codaPkg=FALSE, OpenBUGS.pgm=NULL, working.directory=NULL, clearWD=FALSE, useWINE=FALSE, WINE=NULL, newWINE=TRUE, WINEPATH=NULL, bugs.seed=1, summary.only=FALSE, save.history=(.Platform$OS.type == "windows" | useWINE==TRUE), over.relax = FALSE)
data: either a named list (names corresponding to variable names in the model.file) of the data for the OpenBUGS model, or a vector or list of the names of the data objects used by the model. If data is a one element character vector (such as "data.txt"), it is assumed that data have already been written to the working directory into that file, e.g. by the function bugs.data.
inits: a list with n.chains elements; each element of the list is itself a list of starting values for the OpenBUGS model, or a function creating (possibly random) initial values. Alternatively, if inits=NULL, initial values are generated by OpenBUGS. If inits is a character vector with n.chains elements, it is assumed that inits have already been written to the working directory into those files, e.g. by the function bugs.inits.
parameters.to.save: character vector of the names of the parameters to save which should be monitored
model.file: File containing the model written in OpenBUGS code. The extension must be .txt . The default location is given by working.directory. The old convention allowing model.file to be named .bug has been eliminated because the new OpenBUGS feature that allows the program image to be saved and later restarted uses the .bug extension for the saved images. Alternatively, model.file can be an R function that contains a BUGS model that is written to a temporary model file (see tempfile) using write.model.
n.chains: number of Markov chains (default: 3)
n.iter: number of total iterations per chain (including burn in; default: 2000)
n.burnin: length of burn in, i.e. number of iterations to discard at the beginning. Default is n.iter/2, that is, discarding the first half of the simulations.
n.thin: Thinning rate. Must be a positive integer. The default is n.thin = 1. The thinning is implemented in the OpenBUGS update phase, so thinned samples are never stored, and they are not counted in n.burnin or n.iter. Setting n.thin=2, doubles the number of iterations OpenBUGS performs, but does not change n.iter or n.burnin. Thinning implemented in this manner is not captured in summaries created by packages such as coda.
saveExec: If TRUE, a re-startable image of the OpenBUGS execution is saved with basename (model.file) and extension .bug in the working directory, which must be specified. The .bug files can be large, so users should monitor them carefully and remove them when not needed.
restart: If TRUE, execution resumes with the final status from the previous execution stored in the .bug file in the working directory. If n.burnin=0,additional iterations are performed and all iterations since the previous burnin are used (including those from past executions). If n.burnin>0, a new burnin is performed, and the previous iterations are discarded, but execution continues from the status at the end of the previous execution. When restart=TRUE, only n.burnin, n.iter, and saveExec inputs should be changed from the call creating the .bug file, otherwise failed or erratic results may be produced. Note the default has n.burnin>0.
debug: if FALSE (default), OpenBUGS is closed automatically when the script has finished running, otherwise OpenBUGS remains open for further investigation. The debug option is not available for linux execution.
DIC: logical; if TRUE (default), compute deviance, pD, and DIC. The results are extracted directly from the OpenBUGS log, which uses the rule pD = Dbar - Dhat. If extraction fails or if there are less iterations than required for the adaptive phase, the rule pD=var(deviance) / 2
is computed in R. See bugs.log for more information on extracting results from the log file.
digits: number of significant digits used for OpenBUGS input, see formatC
codaPkg: logical; if FALSE (default) a bugs object is returned, if TRUE file names of OpenBUGS output are returned for easy access by the coda package through function read.bugs. A bugs object can be converted to an mcmc.list object as used by the coda package with the method as.mcmc.list
(for which a method is provided by R2OpenBUGS).
OpenBUGS.pgm: For Windows or WINE execution, the full path to the OpenBUGS executable. For linux execution, the full path to the OpenBUGS executable or shell script (the path to the shell script is not required if the OpenBUGS shell script is in the user's PATH variable). If NULL (unset) and the environment variable OpenBUGS_PATH is set the latter will be used as the default. If NULL (unset), the environment variable OpenBUGS_PATH is unset and the global option R2OpenBUGS.pgm is not NULL the latter will be used as the default. If none of the former are set and OS is Windows, the most recent OpenBUGS version registered in the Windows registry will be used as the default. For other operating systems, the location is guessed by Sys.which("OpenBUGS").
working.directory: sets working directory during execution of this function; OpenBUGS' input and output will be stored in this directory; if NULL, a temporary working directory via tempdir is used.
clearWD: logical; indicating whether the files data.txt , inits[1:n.chains].txt , log.odc , codaIndex.txt , and coda[1:nchains].txt should be removed after OpenBUGS has finished. If set to TRUE, this argument is only respected if codaPkg=FALSE.
useWINE: logical; attempt to use the Wine emulator to run OpenBUGS. Default is FALSE. If WINE is used, the arguments OpenBUGS.pgm and working.directory must be given in form of Linux paths rather than Windows paths (if not NULL).
WINE: Character, path to wine binary file, it is tried hard (by a guess and the utilities which and locate) to get the information automatically if not given.
newWINE: Use new versions of Wine that have winepath
utility
WINEPATH: Character, path to winepath binary file, it is tried hard (by a guess and the utilities which and locate) to get the information automatically if not given.
bugs.seed: Random seed for OpenBUGS. Must be an integer between 1-14. Seed specification changed between WinBUGS and OpenBUGS; see the OpenBUGS documentation for details.
summary.only: If TRUE, only a parameter summary for very quick analyses is given, temporary created files are not removed in that case.
save.history: If TRUE (the default), trace plots are generated at the end.
over.relax: If TRUE, over-relaxed form of MCMC is used if available from OpenBUGS.
To run:
BUGS model in an ASCII file (hint: use write.model).bugs function and run it (see Example section).OpenBUGS window will pop up and will freeze up. The model will now run in OpenBUGS. It might take awhile. You will see things happening in the Log window within OpenBUGS. When OpenBUGS is done, its window will close and will work again.debug=TRUE.BUGS version support:
OpenBUGS>=3.2.1Operation system support:
MS Windowsno problemLinux, intel processorsGUI display and graphics not available.Mac OS X and Unix in generalpossible with Wine emulation via useWINE=TRUEIf useWINE=TRUE is used, all paths (such as
working.directory and model.file, must be given in
native (Unix) style, but OpenBUGS.pgm can be given in
Windows path style (e.g. c:/Program Files/OpenBUGS/ ) or native (Unix) style
(e.g. /path/to/wine/folder/dosdevices/c:/ProgramFiles/OpenBUGS/OpenBUGS321/OpenBUGS.exe ).
If codaPkg=TRUE the returned values are the names of coda output files written by OpenBUGS containing the Markov Chain Monte Carlo output in the CODA format. This is useful for direct access with read.bugs.
If codaPkg=FALSE, the following values are returned: - n.chains: see Section Arguments
n.iter: see Section Arguments
n.burnin: see Section Arguments
n.thin: see Section Arguments
n.keep: number of iterations kept per chain (equal to (n.iter-n.burnin) / n.thin)
n.sims: number of posterior simulations (equal to n.chains * n.keep)
sims.array: 3-way array of simulation output, with dimensions n.keep, n.chains, and length of combined parameter vector
sims.list: list of simulated parameters: for each scalar parameter, a vector of length n.sims for each vector parameter, a 2-way array of simulations, for each matrix parameter, a 3-way array of simulations, etc. (for convenience, the n.keep*n.chains simulations in sims.matrix and sims.list (but NOT sims.array) have been randomly permuted)
sims.matrix: matrix of simulation output, with n.chains*n.keep rows and one column for each element of each saved parameter (for convenience, the n.keep*n.chains
simulations in sims.matrix and sims.list (but NOT sims.array) have been randomly permuted)
summary: summary statistics and convergence information for each saved parameter.
mean: a list of the estimated parameter means
sd: a list of the estimated parameter standard deviations
median: a list of the estimated parameter medians
root.short: names of argument parameters.to.save and deviance
long.short: indexes; programming stuff
dimension.short: dimension of indexes.short
indexes.short: indexes of root.short
last.values: list of simulations from the most recent iteration; they can be used as starting points if you wish to run OpenBUGS for further iterations
pD: an estimate of the effective number of parameters, for calculations see the section Arguments .
DIC: mean(deviance) + pD
Gelman, A., Carlin, J.B., Stern, H.S., Rubin, D.B. (2003): Bayesian Data Analysis, 2nd edition, CRC Press.
Sturtz, S., Ligges, U., Gelman, A. (2005): R2WinBUGS: A Package for Running WinBUGS from R. Journal of Statistical Software 12(3), 1-16.
Andrew Gelman, gelman@stat.columbia.edu ; modifications and packaged by Sibylle Sturtz, sturtz@statistik.tu-dortmund.de , Uwe Ligges, and Neal Thomas
print.bugs, plot.bugs, as well as coda and BRugs packages
## Not run: # An example model file is given in: model.file <- system.file(package="R2OpenBUGS", "model", "schools.txt") # Let's take a look: #file.show(model.file) # Some example data (see ?schools for details): data(schools) schools J <- nrow(schools) y <- schools$estimate sigma.y <- schools$sd data <- list ("J", "y", "sigma.y") inits <- function(){ list(theta=rnorm(J, 0, 100), mu.theta=rnorm(1, 0, 100), sigma.theta=runif(1, 0, 100)) } ## or alternatively something like: # inits <- list( # list(theta=rnorm(J, 0, 90), mu.theta=rnorm(1, 0, 90), # sigma.theta=runif(1, 0, 90)), # list(theta=rnorm(J, 0, 100), mu.theta=rnorm(1, 0, 100), # sigma.theta=runif(1, 0, 100)) # list(theta=rnorm(J, 0, 110), mu.theta=rnorm(1, 0, 110), # sigma.theta=runif(1, 0, 110))) parameters <- c("theta", "mu.theta", "sigma.theta") ## You may need to specify "OpenBUGS.pgm" ## also you need write access in the working directory: schools.sim <- bugs(data, inits, parameters, model.file, n.chains=3, n.iter=5000) print(schools.sim) plot(schools.sim) ## End(Not run)
Useful links