RMark 2.0.8 beta

[jlaake]

If anyone wants to try out the new version of RMark it is at

https://github.com/downloads/jlaake/RMark/RMark_2.0.8.zip

Please let me know if you have any issues. On an unrelated note, if anyone has an example of a Robust Design Barker model that I can use to test addition of that model, please contact me.

I'm working on RMark fairly actively so it will change often until I release it. I have added about 20 more models including 2 species occupancy models, multiple scale occupancy, and most of the CRDMS and ORDMS variants. I've changed the way I add models to RMark which makes it much easier to add them. It now uses the files models.txt and parameters.txt which you'll find in the RMark package directory where you have your R library directory. The same location contains MarkModels.pdf which shows which MARK models are currently supported in RMark. You can look at the text files in Excel (they are tab-delimited) to see the models and parameter names and definitions. Don't edit the file. One of the aspects in RMark that is not explained well in the documentation is the share argument. If you look at parameters.txt you'll see that some parameter pairs are identified like p and c. The first parameter in the pair is the dominant parameter and it is the parameter specification that gets the share=TRUE or FALSE. It is only on those parameters that you can use the share argument and only the pairs designated in the file can be shared. This version now allows epsilon and gamma to be shared in the RDOccupancy model. See ?RDOccupancy for an example.

RMark now has a Namespace which all packages in R are getting. What this means is that not all functions are public and directly accessible. I think I have made all of the common ones public but if I have missed some let me know. You'll know because it will say that it can't find the function. You can get access to any non-public (not exported) function in RMark by using RMark:::function. That works with any R package. But I'd just as soon have any common function exported.

regards --jeff

[Roman Luštrik]

For my own needs, I modified the extract.mark.output function to extract some extra parameters from the .out file (Estimated Derived Parameters as I couldn't find a function that would give me this for each model I have) and now setup.model is caged inside a namespace. Although I've solved this by calling the function specifically out using the :::, it would perhaps be nice to have it out in the open. Or even better, have a function that would scrape the entire .out file. If you have a few example .out files, I would love to lend a hand to write a scraper function that would be general enough to work for all models.

Cheers,
Roman

[jlaake]

Roman-

The derived parameters should be extracted from the binary file. Which models were not getting the derived parameters? What other portions of the .out do you need that are not now extracted?

--jeff

[Roman Luštrik]

I'm doing a Huggins closed captures model. For the life of me, I couldn't find these parameters anywhere in the output of extract.mark.output.

Code: Select all

                     Estimates of Derived Parameters
                   Population Estimates of { p(~1)c() }
                                                95% Confidence Interval
     Group     N-hat        Standard Error      Lower           Upper
 ---------  --------------  --------------  --------------  --------------
       1    675.23114       16.925849       646.17133       712.98151 

[jlaake]

Look at model$results$derived. See example below from ?edwards.eberhardt.The derived parameters are extracted from the binary file with the function read.mark.binary ( or read.mark.binary.linux for linux machines). I have had problems with getting derived parameters from pradel and popan. I created popan.derived to create the popan derived parameters because I and others needed them. The formats vary in the output so the binary file is a better place to extract them. I need to work out with Gary why binary file for some model types have not worked for me. --jeff


> ee.results[[10]]$results$derived
estimate se lcl ucl
1 95.97188 6.702416 86.52793 113.8874

[Roman Luštrik]

Looks like I didn't look hard enough. Worked like a charm, thanks!