Hello everyone,
I just uploaded a new version of
unmarked to CRAN. The windows and mac binaries should be ready within a few days. Release notes can be found here:
https://sites.google.com/site/unmarkedinfo/. There are two important changes that I want to highlight.
First, we added the remarkable model of Dail and Madsen (Biometrics
early view), which relaxes the closure assumption of the N-mixture model and allows for inference about the dynamics of open populations. See ?pcountOpen for more details.
Second, unmarked now depends on the
Rcpp and
RcppArmadillo packages. Once the binaries are ready, you can get everything at once by typing:
install.packages("unmarked", dependencies=TRUE)
These packages make it easy to call C++ code from R. Currently we are only making slight use of these packages, but eventually we hope to see major speed improvements by writing all the likelihoods in C++.
Comments and feedback are welcome as always.
Richard