BookmarkSubscribeRSS Feed
dwitiyas
Calcite | Level 5

How do I calculated S(t) at a particular time t using proc phreg.

Proc phreg data = a.input;

model tenure * target (0) =a1 a2 a3/ selection = stepwise ties = efron;

baseline out = out1 covariates = sample survival = s;

run;

I the above example if I pass covariates = sample it gives me survival estimates for all possible values of t present in a.input. My datasets are very huge and I am getting memory error when I try to score. I need to calculated s(30) only for my dataset "sample".

6 REPLIES 6
Reeza
Super User

Timelist option on the baseline statement. See the docs.

SAS/STAT(R) 9.2 User's Guide, Second Edition

dwitiyas
Calcite | Level 5

Isn't this option only for bayesian model.  I am doing a classical model here.

Reeza
Super User

If you run out of memory perhaps scoring by 'hand', a manual calculation in datastep?

dwitiyas
Calcite | Level 5

Yes but I need help on this. I understand I can get Xbeta value and the baseline estimates. I need help with the formula to maunally calc it . My values are not matching with what sas provides.

Reeza
Super User

I'm sure there's many ways to do that, but I don't feel like breaking out my stats book Smiley Happy

How about breaking up your sample dataset and running the whole thing a few times via a macro.

dwitiyas
Calcite | Level 5

That is what I am doing right now Smiley Happy If I get the manual calc for survival estimates I would definitely go that way.

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

How to choose a machine learning algorithm

Use this tutorial as a handy guide to weigh the pros and cons of these commonly used machine learning algorithms.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 6 replies
  • 1065 views
  • 0 likes
  • 2 in conversation