topic Re: PROC GLMMIX did not converge even with only one binary covariate in Statistical Procedures

PROC GLMMIX did not converge even with only one binary covariate

lichee — Tue, 29 Aug 2023 21:12:23 GMT

Hi SAS Gurus,

I am trying to run HGLM with PROC GLIMMIX using the data of around six million of individuals across states with state effect. Even though I started with the simplest model with only one covariate included only, the model did not converge. Any of you have any insight or suggestion?

Below is the log:

Any states are excluded when they had extremely low counts of individuals with the outcome variable flag_success_discharge=1.

Thank you,

Lichee

Re: PROC GLMMIX did not converge even with only one binary covariate

PaigeMiller — Tue, 29 Aug 2023 21:19:53 GMT

Plenty of discussion about GLIMMIX not converging here in the forums. One example: https://communities.sas.com/t5/Statistical-Procedures/Glimmix-did-not-converge/td-p/387883 You should search for other discussions.

Re: PROC GLMMIX did not converge even with only one binary covariate

jiltao — Wed, 30 Aug 2023 01:23:39 GMT

What happens if you add METHOD=LAPLACE option in the PROC GLIMMIX statement?

Also, you might want to spell out: dist=binary in the MODEL statement.

Thanks,

Jill

Re: PROC GLMMIX did not converge even with only one binary covariate

lichee — Wed, 30 Aug 2023 02:05:27 GMT

Thank you!

I added METHOD=LAPLACE, and now the initial estimates did not yield a valid objective function.

Re: PROC GLMMIX did not converge even with only one binary covariate

Ksharp — Wed, 30 Aug 2023 11:51:52 GMT

If I was right, your response variable "flag_success_discharge" is continuous variable,
NOT a category variable. Use PROC FREQ to check its levels firstly.

Re: PROC GLMMIX did not converge even with only one binary covariate

lichee — Wed, 30 Aug 2023 12:04:15 GMT

Thank you! The dependent variable is dichotomous. Below is the crosstab:

Re: PROC GLMMIX did not converge even with only one binary covariate

Ksharp — Wed, 30 Aug 2023 12:19:01 GMT

Try specify the distribution :
/d=b
====>
/dist=binary

Re: PROC GLMMIX did not converge even with only one binary covariate

lichee — Wed, 30 Aug 2023 13:13:13 GMT

Yes, I re-ran the code with dist=binary specified and got the same message "Did not converge"

Re: PROC GLMMIX did not converge even with only one binary covariate

jiltao — Wed, 30 Aug 2023 13:20:08 GMT

Try using PARMS statement to enter you own starting values. This might take several trial and error.

If that does not help, can you send in your data?

Thanks,

Jill

Re: PROC GLMMIX did not converge even with only one binary covariate

SteveDenham — Wed, 30 Aug 2023 19:30:49 GMT

Try increasing the number of iterations in an NLOPTIONS statement. Binary data often require more than the default 20 iterations.

SteveDenham

Re: PROC GLMMIX did not converge even with only one binary covariate

Ksharp — Thu, 31 Aug 2023 11:52:49 GMT

@Rick_SAS 's blog could give you a clue.

https://blogs.sas.com/content/iml/2019/04/03/g-matrix-is-not-positive-definite.html