write_workflow {Infusion} | R Documentation |
Workflow template
Description
codewrite_workflow writes a workflow script for inference. Beyond possibly saving some typing, this suggests what may be a reasonably good starting workflow. One should not expect to control all options of the workflow through the write_workflow
arguments.
Usage
write_workflow(con = stdout(), lower, upper, nUnique, Simulate, simulator_args=NULL, ...)
Arguments
con |
A connection object or a character string. Passed to |
lower |
A named numeric vector of lower bounds for parameter space. |
upper |
A named numeric vector of upper bounds for parameter space. |
nUnique |
Number of simulations of the process (i.e. of rows of the reference table) in the first iteration. |
Simulate |
Function, or function name as a string. Sets the same-named |
simulator_args |
list of arguments for the simulator. Sets the ... in the |
... |
Sets additional arguments in the |
Value
No return value. Used for the side-effect text, written to the connection.
Examples
write_workflow(
## arguments for init_reftable():
lower=c(logTh1=-2,logTh2=-2,logTh3=-2,logTh4=-2,ar=0.01,logMu=-5,MEANP=0.01),
upper=c(logTh1=1, logTh2=1, logTh3=1, logTh4=1, ar=0.99,logMu=-2,MEANP=0.99),
nUnique = 1000,
#
## for add_reftable():
Simulate="schtroumf", # name of a user-defined R function
simulator_args= list( # Imagine that schtroumf() require arguments 'exe_path' and 'cmdline':
exe_path="'path_to_smurf_executable'",
cmdline="'smurf.exe -a -b -c -d'"
), # Do check the quotation marks in the output...
#
## optional arguments for refine():
n=8000/3.2, CIs=TRUE, update_projectors=FALSE)