Prev: Proc mianalyze and nonmissing data?
Next: Help w/ Difference-in-Difference
From: machelle on 14 Apr 2010 13:07
HI there
Does anyone know what to do when confronted with an "out of memory"
error?
This happened when trying to run a pooled logistic regression GEE
model with an unstructured correlation matrix. Note that the same
model using either the independent or exchangeable correlation matrix
structure doesn't have this problem.

Here's the code from my SAS Log:

24 /*1.2 Unstructured correlation matrix*/
25 title '1.3 hospital cluster TDCOX using Unstructured correlation
matrix';
26 proc genmod data=Lib1.stopExcl_Resp10b descending;
27 class idkey HOSP;
28 model FirstResp= StopBFExcl male SomeUniversity
NoneBF_gt_3mo CSection complication
28 ! GestAge
29 Smoked ATOPIC_H AGE_MC hospnonBM_score
Low_Birthwaz
30 IncDeltaWaz ALC_MLperMonth Rash TDC_OthIll
TDC_Resp_Hosp
31 Time Time1-Time2/ link=logit dist=bin;
32 repeated subject=HOSP/ type=Un;
33 run;

NOTE: Class levels for some variables were not printed due to
excessive size.
NOTE: PROC GENMOD is modeling the probability that FirstResp='1'.
NOTE: Algorithm converged.
ERROR: Out of memory.
NOTE: The SAS System stopped processing this step because of
insufficient memory.
NOTE: PROCEDURE GENMOD used (Total process time):
real time 12.82 seconds
cpu time 3.43 seconds

Any ideas?

thanks,
Machelle

 | 
Pages: 1
Prev: Proc mianalyze and nonmissing data?
Next: Help w/ Difference-in-Difference