search

dswah / pyGAM

[HELP REQUESTED] Generalized Additive Models in Python
https://pygam.readthedocs.io
Apache License 2.0
862 stars 159 forks source link

LinearGAM() vs. GAM()? #233

Closed moptis closed 5 years ago

moptis commented 5 years ago

Perhaps not the best place for this question, but I was curious what the difference is. Are both suitable for regression?

arose13 commented 5 years ago

GAM() is general and can be used for anything, It defaults to distribution='normal', link='identity' which makes it identical to LinearGAM(). In fact LinearGAM etc all inherits the GAM() class.

Check out the documentation for more information birb docs

dswah commented 5 years ago

@moptis sorry for the poor response time. @arose13 is absolutely right; the GAM() class is the general class from which LinearGAM and the rest inherit.

This means that GAM typically lacks some convenient functionality specific to the other models and distributions. For example LinearGAM has a prediction intervals method which the rest do not currently have.