09-09-2014 11:04 AM
I initially built a logistic model for my dataset, and used the score option within proc logistic to get a probability of the event for each observation. I would like to use proc genmod (with binomial distribution, logit link) so I can use the bayes option with coeffprior= in order to get coefficients based on a categorical variable.
The issue - I don't know how to get the probability of the event for each observation within proc genmod. How can I do something like the score option within proc logistic?
09-09-2014 11:26 AM
You need to use ods with obstats option something liket this.
ods output obstats=obstats;
PROC GENMOD DATA=have;
MODEL Y=X / D=B OBSTATS;
proc print data=obstats;
09-09-2014 01:25 PM
Thank you for your response. Gave that a try, but the proc print tells me there's no dataset called obstats. In the line: ods output obstats=obstats, the obstats does not become blue as a recognized option.
09-09-2014 01:58 PM
Yes, I put that in. Actually, I just realized I got a note too
NOTE: The OBSTATS, RESIDUALS, XVARS, PREDICTED, and DIAGNOSTICS options are ignored for a Bayesian analysis.