Valid character strings for different optimization routines

[tommyg]

What are the valid character strings for the options argument in function mark ?

Code: Select all

mark(data.proc, data.ddl, options="?", . .

Supplying a nonsense character string did not return any warning, nor a list of the valid options. I believe simulated annealing and MCMC are the only alternatives to the default Newton-Raphson routine, but I don't know what string would follow the proc estimate statement in the *inp file mostly because I started using RMark before I started using Mark.

Thanks.

[jlaake]

So you want to know all the secret passwords to the adventure game, eh!? Just kidding. I honestly don't know them all and the only 2 I know are documented in the help.

\code{options} is a character string that is tacked onto the \code{Proc Estimate} statement for the MARK .inp file. It
can be used to request options such as NoStandDM (to not standardize the design matrix) or SIMANNEAL (to request use of the simulated annealing optimization method) or any existing or new options that can be set on the estimate proc.

I wrote that help statement generically because it can be used that way but I don't know all the other options. The inp file is not something that a MARK user manipulates directly and it is constructed by the MARK interface based on options selected in the interface. Other than simulated annealing I've never had any reason to use other options but then again I don't know what exists for those options. I'd have to try all the various choices in MARK interface and look at what it created in the .inp file. Because Gary doesn't expect the user to ever see the .inp file he doesn't document it in the help in that way. Probably not the answer you were looking for but that's it.

MCMC cannot be handled that way. I have plans to include it but that will require some coding for all of the input needed for MCMC.

regards --jeff

[bacollier]

TommyG,

I don't know if there is a list as I did a quick look and I could not find one, but I don't think MCMC is one of the options. I could hazard a guess that some of the statements are the same ones Gary used in SURVIV, which I suspect he has coded into MARK elsewhere now.

Bret

[jlaake]

It is often hard to sift through all of the documentation. One useful tip is to download the pdf version of the help files

http://cran.r-project.org/web/packages/RMark/RMark.pdf

For each package on CRAN you'll find a pdf. The search facility in R of package help is lacking but you can use ctrl-f in adobe reader for any pdf to search for strings. That will work with all of the RMark documentation sources: Appendix C, workshop notes, and the help file in pdf format.

--jeff