Usage
autorun.jags(model, monitor = NA, data=NA, n.chains=NA,
inits = NA, startburnin = 4000, startsample = 10000,
datalist=NA, initlist=NA, psrf.target = 1.05, normalise.mcmc = TRUE,
check.stochastic = TRUE, modules=runjags.getOption('modules'),
factories=runjags.getOption('factories'),
raftery.options = list(), crash.retry=1, summarise = TRUE,
confidence=0.95, plots = runjags.getOption('predraw.plots') && summarise,
thin.sample = FALSE, jags = runjags.getOption('jagspath'),
silent.jags = runjags.getOption("silent.jags"), interactive=FALSE,
max.time=Inf, adaptive=1000, thin = 1, monitor.deviance = FALSE,
monitor.pd = FALSE, tempdir=runjags.getOption('tempdir'),
jags.refresh=0.1, batch.jags=silent.jags,
method=runjags.getOption('method'), method.options=list())autoextend.jags(runjags.object, add.monitor=character(0),
drop.monitor=character(0), drop.chain=numeric(0),
combine=length(c(add.monitor,drop.monitor,drop.chain))==0,
startburnin = 0, startsample = 10000, psrf.target = 1.05,
normalise.mcmc = TRUE, check.stochastic = TRUE,
raftery.options = list(), crash.retry=1, summarise = TRUE,
confidence=0.95, plots = runjags.getOption('predraw.plots') && summarise,
thin.sample = FALSE, jags = runjags.getOption('jagspath'),
silent.jags = runjags.getOption('silent.jags'), interactive=FALSE,
max.time=Inf, adaptive=1000, thin = runjags.object$thin,
tempdir=runjags.getOption('tempdir'), jags.refresh=0.1,
batch.jags=silent.jags, method=NA, method.options=NA)
Arguments
model
either a relative or absolute path to a textfile (including
the file extension) containing a model in the JAGS language and possibly
monitored variable names, data and/or initial values, or a character
string of the same. No default. The model must be s
monitor
a character vector of the names of variables to monitor.
The special node names 'deviance', 'pd', 'pd.i', 'popt' and 'dic' are
used to monitor these model fit diagnostics (see the JAGS user manual
for more information), but with the exception of 'devianc
data
either a named list or a character string in the R dump
format containing the data. If left as NA, the model will be run
without external data.
n.chains
the number of chains to use with the simulation. More
chains will improve the sensitivity of the convergence diagnostic, but
will cause the simulation to run more slowly (although this may be
improved by using a method such as 'parallel' or 'snow'). The
inits
either a character vector with length equal to the number
of chains the model will be run using, or a list of named lists
representing names and corresponding values of inits for each chain, or
a function with either 1 argument representing the chain or
runjags.object
the model to be extended - the output of a
run.jags (or autorun.jags or extend.jags etc) function, with class
'runjags'. No default.
add.monitor
a character vector of variables to add to the
monitored variable list. All previously monitored variables are
automatically included - although see the 'drop.monitor' argument.
Default no additional monitors.
drop.monitor
a character vector of previously monitored variables
to remove from the monitored variable list for the extended model.
Default none.
drop.chain
a numeric vector of chains to remove from the extended
model. Default none.
combine
a logical flag indicating if results from the new JAGS
run should be combined with the previous chains. Default TRUE if not
adding or removing variables or chains, and FALSE otherwise.
startburnin
the number of burnin iterations, NOT including the
adaptive iterations to use for the initial pilot run of the chains.
startsample
the total number of samples (including the chains
supplied in runjags.object for autoextend.jags) on which to assess
convergence. If the runjags.object already contains this number of
samples then convergence will be assessed on this object, otherwise th
datalist
an optional named list containing variables used as
data, or alternatively a function (with no arguments) that returns a
named list. If any variables are specified in the model block using
'#data# ', the value for the corresponding named variab
initlist
an optional named list containing variables used as
initial values, or alternatively a function (with a single argument
representing the chain number) that returns a named list. If any
variables are specified in the model block using '#inits# '
psrf.target
the value of the point estimate for the potential
scale reduction factor of the Gelman Rubin statistic below which the
chains are deemed to have converged (must be greater than 1). Default
1.05.
normalise.mcmc
the Gelman Rubin statistic is based on the
assumption that the posterior distribution of monitored variables is
roughly normal. For very skewed posterior distributions, it may help to
log/logit transform the posterior before calculating the Gelman Rubin
check.stochastic
non-stochastic monitored variables will cause
errors when calculating the Gelman-Rubin statistic, if
check.stochastic==TRUE then all monitored variables will be checked to
ensure they are stochastic beforehand. This has a small computational
cost, which
modules
a character vector of external modules to be loaded into
JAGS. More than 1 module can be used. Default none.
factories
a character vector of factory modules to be loaded into
JAGS. More than 1 factory can be used. Factories should be provided in
the format '()', for example:
factories='mix::TemperedMix(sampler)'. Also ensure that any required
modules
raftery.options
a named list which is passed as additional
arguments to raftery.diag. Default none (default
arguments to raftery.diag are used). crash.retry
the number of times to re-attempt a simulation if the
model returns an error. Default 1 retry (simulation will be aborted
after the second crash).
summarise
should summary statistics be automatically calculated
for the output chains? Default TRUE.
confidence
the prob argument to be passed to HPDinterval for
calculation of confidence intervals. Default 0.95 (95% confidence
intervals).
plots
should traceplots and density plots be pre-drawn by runjags
to facilitate more convinient assessment of convergence after the model
has finished running? If TRUE, the returned list will include elements
'trace' and 'density' which consist of a list of la
thin.sample
option to thin the final MCMC chain(s) before
calculating summary statistics and returning the chains. Thinning very
long chains allows summary statistics to be calculated more quickly. If
TRUE, the chain is thinned to as close to a minimum of startsamp
jags
the system call or path for activating JAGS. Default calls
findjags() to attempt to locate JAGS on your system.
silent.jags
should the JAGS output be suppressed? (logical) If
TRUE, no indication of the progress of individual models is supplied.
Note that output will still be produced by runjags even if silent.jags
is set to TRUE - to suppress all output set silent.jags and si
interactive
option to allow the simulation to be interactive, in
which case the user is asked if the simulation should be extended when
run length and convergence calculations are performed and the extended
simulation will take more than 1 minute. The function will
max.time
the maximum time for which the function is allowed to
extend the chains to improve convergence, as a character string
including units or as an integer in which case units are taken as
seconds. Ignored if interactive==TRUE. If the function thinks that th
adaptive
the length of the adaptive phase to use when re-compiling
models. This will be run for every new simulation except for the rjags
method where it is only required to be run during the model compilation
phase (this will only be performed once). Note that
thin
the thinning interval to be used in JAGS. Increasing the
thinning interval may reduce autocorrelation, and therefore reduce the
number of samples required, but will increase the time required to run
the simulation. Using this option thinning is performe
monitor.deviance
this argument is deprecated and remains for
backwards compatibility only. See the 'monitor' variable.
monitor.pd
this argument is deprecated and remains for backwards
compatibility only. See the 'monitor' variable.
tempdir
option to use the temporary directory as specified by the
system rather than creating files in the working directory. Any files
created in the temporary directory are removed when the function exits
for any reason. Default TRUE.
jags.refresh
the refresh interval (in seconds) for monitoring
JAGS output using the 'interactive' and 'parallel' methods (see the
'method' argument). Longer refresh intervals will use less processor
time. Default 0.1 seconds.
batch.jags
option to call JAGS in batch mode, rather than using
input redirection. On JAGS >= 3.0.0, this suppresses output of the
status which may be useful in some situations. Default TRUE if
silent.jags is TRUE, or FALSE otherwise.
method
the method with which to call JAGS; probably a character
vector specifying one of 'rjags', 'simple', 'interruptible', 'parallel',
'rjparallel' or 'snow' (and see also xgrid.autoextend.jags).
method.options
an optional named list of argument to be passed to
the method function (including a user specified method function). Of the
default arguments, only 'nsims' indicating the number of separate
simulations (for parallel, snow and bgparallel methods) and 'cl'