BookmarkSubscribeRSS Feed
0 Likes

I do a fair bit of work with ecological datasets trying to estimate population size using techniques like mark-repature modeling and occupancy estimation. I think it would be great if SAS has a proceedure for some of these types of models that did not require coding them up in PROC NLIN or PROC NLMIXED. Anyone else out there had a similar experience?

4 Comments
PGStats
Opal | Level 21

For nonlinear modelling and simulation, it is also worth looking at proc model (SAS/ETS) and the Nonlinear Programming Solver of proc optmodel (SAS/OR).

brooksj029
Fluorite | Level 6
Thank you for the suggestion PGStats. The MODEL procedure may be simpler
than NLIN. However, given the large number of models I run and the
complexity of the models, writing the code is still labor intensive. I was
imagining a procedure with an ease of use similar to GLMMIX or GENMOD.

##- Please type your reply above this line. Simple formatting, no
attachments. -##
SASCom1
SAS Employee

SAS does not have procedures dedicated specifically to analysis of capture-recapture data.  However, existing procedures may be useful.  For example, log-linear models can be used with such data for the problem of finite population size estimation.  Such models can be fit as poisson regression models using PROC GENMOD.  Examples of using PROC GENMOD for fitting such models with this sort of data are given in the book "Advanced Log-Linear Models Using SAS" (Zelterman, D., SAS Institute, 2002).  You can order this book from SAS web site,

 

http://support.sas.com/publishing/index.html

 

or call 1-800-727-3228.

brooksj029
Fluorite | Level 6
Thank you for the suggestion SASCom1. I will check this out.

##- Please type your reply above this line. Simple formatting, no
attachments. -##