topic Re: proc glimmix choose different estimation methods in SAS Programming

proc glimmix choose different estimation methods

ANKH1 — Tue, 06 Feb 2024 14:40:22 GMT

In proc glimmix, what syntax is used to change the optimization algorithm to other than Newton Raphson?

proc glimmix data=mydata;
   class myclass;
   model response = mycovariates / dist=binomial link=logit solution;
   random intercept / subject=myid;
 run;

Re: proc glimmix choose different estimation methods

sbxkoenk — Tue, 06 Feb 2024 18:18:58 GMT

SAS® 9.4 and SAS® Viya® 3.5 Programming Documentation
SAS/STAT 15.3 User's Guide
The GLIMMIX Procedure
NLOPTIONS Statement
https://go.documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/statug/statug_glimmix_syntax18.htm

SAS/STAT 15.3 User's Guide
Shared Concepts and Topics
NLOPTIONS Statement
https://go.documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/statug/statug_introcom_sect052.htm

Koen

Re: proc glimmix choose different estimation methods

ANKH1 — Mon, 12 Feb 2024 17:08:16 GMT

proc glimmix data=mydata method=laplace;
   class myclass;
   model response = mycovariates / dist=binomial link=logit solution;
   random intercept / subject=myid;
 run;

Hi! I found out that you can choose the method (example laplace) by adding the statement above.

Re: proc glimmix choose different estimation methods

sbxkoenk — Mon, 12 Feb 2024 17:29:00 GMT

@ANKH1 wrote:
proc glimmix data=mydata method=laplace;
run;
Hi! I found out that you can choose the method (example laplace) by adding the statement above.

There's an "inter-play" between several statements and options.

Several estimation methods of the GLIMMIX procedure (METHOD=RSPL, MSPL, RMPL, MMPL) are doubly iterative.
The NLOPTIONS statement controls the inner iterations. The outer iteration behavior can be controlled with options in the PROC GLIMMIX statement, such as the MAXLMMUPDATE=, PCONV=, and ABSPCONV= options. If the estimation method involves a singly iterative approach, then there is no need for the outer cycling and the model is fit in a single optimization controlled by the NLOPTIONS statement (see the section Singly or Doubly Iterative Fitting).
Note that in a GLMM with pseudo-likelihood estimation, specifying TECHNIQUE=NONE has the same effect as specifying the NOITER option in the PARMS statement. If you estimate the parameters by METHOD=LAPLACE or METHOD=QUAD, TECHNIQUE=NONE applies to the optimization after starting values have been determined.

SAS Help Center

https://go.documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/statug/statug_glimmix_syntax18.htm

Br, Koen