[R] Generalized Additive Models - gamma against overfitting

Mike.lang michael.lang at tum.de
Wed Nov 13 10:03:07 CET 2013

Dear listeners, 

in order to analyze a dataset, which includes more than 30k records, I'm
using a general GAMM like y~s(x)+s(y)+z+e with a underlying normal
distribution. Unfortunately, I have some problems with over- and
underfittings in smooth-functions at the same time. 
In general there is the possibility to set up a gamma like gamma=1.4 in
order to avoid overfitting in the smooth-functions. In my case I would need
different gammas for different smooth-functions, because s(x) is overfitted
and s(y) underfitted currently. 
So, is there any possibility to set up a gamma for each smooth-function

I really would appreciate your responses! 

Thanks a lot!

View this message in context: http://r.789695.n4.nabble.com/Generalized-Additive-Models-gamma-against-overfitting-tp4680364.html
Sent from the R help mailing list archive at Nabble.com.

More information about the R-help mailing list