01-15-2014 04:35 PM
I have a model that has more than 10 random effects identified by bcoef1-bcoef20. Is there a compact way to write the random statement in PROC NLMIXED other than the usual way of:
"random bcoef1, bcoef2, ..., bcoef20 ~normal([0,0,...,0],[vcoef1,
vcoef20]) subject=ID out=outre;"
01-16-2014 02:00 PM
Not that I am aware of. Are you truly trying to estimate 190 covariance parameters in a nonlinear model? If so, I want to know what platform (OS, memory, etc.), so that I can talk our company into buying comparable equipment.
01-16-2014 09:57 PM
Thank you. Unfortunately, we have to run a non linear model with more than 10 random effects. We have not tested this out on our computer yet but if it runs successfully, I will inform you.