[R] Gam (mgcv) function

Cetinyürek Aysun Aysun.Cetinyurek at ulg.ac.be
Wed Aug 19 13:56:49 CEST 2009

Dear all,

I have obtained a model for my data set using gam function in mgcv library.
The model is as follows:

Family        : Binomial
Link function : Logit
RU ~ s(time, by = Status, bs = "ps") + Region*Status+ Gender*Status

I have obtained the results.

While presenting the results for Region and Gender, do I consider the
model as usual logistic and estimate the ORS by exponentiating the
coefficient estimate (and also for 95% CIs of ORs)? or is there any change
in the calculation of OR?
I have been searching for that and the thing I have found was related to
model prediction and plotting of smooth terms.

Thank you in advance,

Kind Regards,


More information about the R-help mailing list