Generalized Additive Models: Allowing for some wiggle room in your models

[@nannangao7256]

Thanks so much for so excellent explanation of GAM

[@brazilfootball]

Does anyone know if backward selection is a viable option with GAMs?

[@ramansingh-cg7rs]

many thanks for posting this video, very helpful, would you know of any C++ implementation of mgcv?

[@convoluted2348]

you should learn python or r it will save you time in the long run

[@piotr780]

we can also use s(x1, x2) alone instead of te(x1, x2) alone, which is surprising: n = 1000 x1 = rnorm(n) x2 = rnorm(n) y = x1 * x2 model = gam(y ~ s(x1, x2 )) summary(model) plot(model) # quite natural - set of rescaled y=1/x functions

[@TheKoekiemonster1234]

S stands for surprising

[@nightcross1030]

The tensor smooth is different from calling `s`, for reasons. Check the documentation.