Proc genmod for train and test subsets

[svzplayer]

Hello everyone,

I have a following problem: I have divided dataset for two subsets - train and test. I wanted to perform three analysis, i.e. simple lin.reg., using box-cox and using genmod with gamma distribution. For the first two I have fitted model for train and calculated it later using proc score.

However, I'm stuck with genmod. How can I predict values (for dataset test) here? Is there any way to calculate it automatically?

Thanks in advance!

[PaigeMiller]

One method is to use the STORE command in PROC GENMOD, and then PROC PLM to score your new test data set.