New RMark Feature

[jlaake]

One of the most frequently asked questions on the RMark subforum has been "how do I predict real values for a range of covariates that are in the design data"? My workaround suggestion has been to code the design covariates as individual covariates with the same value for each critter. If you have a large problem, individual covariates can create very long execution times in MARK. So after many excuses I finally got around to coding a function that will allow predictions for a range of design covariates as well as individual covariates.

This is not ready for prime time so I did not include it in the recent v2.1.14 on CRAN. Instead I have put it in what will eventually be v2.2.0. You can get v2.2.0 from my google drive at https://drive.google.com/open?id=0B77g1ScdUwVeVU9rUE5lVS1kZ2M. Windows users can download the .zip file for 2.2.0 and then use Packages/Install from local zip in R. Linux and Mac users can download the tar.gz file and use install.packages("RMark_2.2.0.tar.gz",type="source")

Once installed, see ?predict_real. I've tested this function under a few conditions but would appreciate help with further testing and comments on the function and documentation.

Feel free to contact me directly with regard to the new function.

[jlaake]

For those of you that may have already downloaded this, I found a bug and fixed it. It didn't work properly when order of variable names in df didn't match order in data. It works now regardless of order. If you already downloaded 2.2.0, please do so again before using predict_real.

--jeff