N.B. This functionality is only provided on request for expert users of external model software. It is assumed that users are capable of setting up batch files, executing them from R, and reading in the results. The authors can only provide limited support.
With the setting control$parallel="snow.chains"
, dream
interprets FUN
to be of the form logp=f(chain.id, list of
pars)
. By passing the chain.id
to the function, the user can
run external programs in parallel (using SNOW) within separate
folders.
In addition:
func.type
must be "logposterior.density",
i.e. FUN
must return the log-likelihood
measurement
should be left at its default value of
NULL
or dream will stop with an error.
FUN.pars
should be left at its
default value of list()
or dream will stop with an error.
dreamCalibrate cannot be used
dream
for documentation of other arguments, and
dreamCalibrate
for the recommended way to calibrate a
function using dream.
demo("parallelisation_chain_id")
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.