admbControl: Control options for ADMB

In bbolker/glmmadmb: Generalized Linear Mixed Models using 'AD Model Builder'

Description

Determines options for running AD Model Builder

Usage

admbControl(impSamp=0,maxfn=500,imaxfn=500,maxph=5,noinit=TRUE,shess=TRUE,
  run=TRUE, ZI_kluge=FALSE, poiss_prob_bound=TRUE)

Arguments

impSamp	Number of importance sampling steps
maxfn	Maximum number of evaluation steps
imaxfn	number of function evaluations used in intermediate optimization steps.
maxph	maximum phase
noinit	Don't initialize random effects before inner optimization: see ADMB-RE reference manual
shess	Use sparse Hessian: see ADMB-RE reference manual
ZI_kluge	retain a zero-inflation term of 0.001 even when zeroInflation=FALSE (for backward compatibility/testing; deprecated)
poiss_prob_bound	add a small term to prevent estimated Poisson probabilities of zero
run	Whether to run ADMB at all. This will write out files to the appropriate working directory, or read existing output files in the working directory (glmmadmb.{std,bar,par,cor}) into a glmmadmb object.

Details

See the AD Model Builder reference manuals at http://admb-project.org/documentation/manuals/admb-user-manuals

Value

Returns a list of options suitable for passing as the admb.opts argument to glmmadmb

Troubleshooting

Some useful tricks for diagnosing and overcoming fitting problems:

• Diagnostics

  • to find the version of glmmADMB you are running, use sessionInfo

  • to find the specific version of the ADMB binaries built into the package, use get_bin_version

  • to see exactly where glmmADMB is looking for the executable on your system, and to see what command-line arguments it is passing to the ADMB executable, use debug=TRUE

  • to see lots more output from ADMB, use verbose=TRUE

  Fixing problems

  • if the fit fails with "couldn't find STD file", try setting noinit=FALSE and/or shess=FALSE

  • if you want to try running the ADMB model outside of R, use run=FALSE and set save.dir in the glmmADMB arguments. This will result in a saved directory containing the glmmadmb executable (on systems other than Windows) as well as glmmadmb.pin and glmmadmb.dat files. You can then run the glmmadmb executable with appropriate command-line arguments in a command shell or terminal window. If you run the same R command again (i.e. still with run=FALSE and with the same save.dir) glmmADMB will find any output files that have been produced and read them into a glmmadmb object. You can also try sending the glmmadmb.pin and glmmadmb.dat files to an ADMB expert for diagnosis.

  • if the ADMB output shows a statement like "Use the [] command line option ...", use the extra.args argument to glmmadmb to set command line options for ADMB. ADMB's diagnostic statements are usually fairly specific about which command line options to use, but see the ADMB reference manuals (linked from Details below) for more information.

Author(s)

Ben Bolker

Examples

admbControl(impSamp=200)
## Not run: 
## running to produce saved files in directory:
data(bacteria,package="MASS")
bacteria$present <- as.numeric(bacteria$y)-1
## run to generate files
glmmadmb(present ~ trt + I(week > 2), random = ~ 1 | ID,
                     family = "binomial", data = bacteria,
         save.dir="tmp",admb.opts=admbControl(run=FALSE),
         debug=TRUE)
## now run glmmadmb outside of R ...
## ./glmmadmb -maxfn 500 -maxph 5 -noinit -shess
## run to read in data
result <- glmmadmb(present ~ trt + I(week > 2), random = ~ 1 | ID,
                     family = "binomial", data = bacteria,
         save.dir="tmp",admb.opts=admbControl(run=FALSE))

## End(Not run)

bbolker/glmmadmb documentation built on May 11, 2019, 9:29 p.m.