08-14-2012 09:33 AM
I am trying to run an analysis for survival analysis with a recurrent event using PROC PHREG.
Though my data set is quite big with many categorical variables some with up to nine levels.
At first I ran the following models I. Count data model ii) conditional model taking possible correlation between events into considerations.
I noticed that it took like forever to run each analysis. Then I removed all covariates but one to check for the proportionality assumption of that particular covariate by add additional line to my code. Each time I run the code, my sas end up stopped working with the message "sas .exe stopped working". The only way to close the application or get it running again is to restart my system.
below is the code that kept crashing my sas 9.2.
PROC PHREG data=myday covs(aggregate) covm;
model(time_start time_stop)*cen(0) = var_a /RL;
proportinality_check=(time_stop - time_start)*var_a;
1. Please does anyone know what I might do to get this fixed.
2. In the event that I get it fix, can I possible in my final model, check the proportionality assumption of all other variables at same time?
Thanks in anticipation of a swift and favorable response.
08-14-2012 09:56 AM
Any time that SAS crashes, you should notify SAS Technical Support so that
1) SAS can fix the problem in a future release, and/or
2) Technical Support can tell you about a workaround, hotfix, or other known solution.
This Web page contains instructions for submitting problems to SAS Technical Support: http://support.sas.com/techsup/contact/