Hi,
I'm working on a model to predict claims (in my case a binary target) on a
sample of policies that have been exposed to risk for varying periods of
time. It seems reasonable to suppose that the claim is proportional to the
time-on-risk. How do I deal with this in Logistic Regression? Should I
include a term such as log(exposure time)?
Thanks for your help!
Regards,
L.
|