[R] bivariate function in gam model

[Stephen Ellner]

I don't think this actually is a GAM model
 >       logit(p) = a + f(x1) + f(x2) + f(x1,x2)
unless you mean f(x1*x2) or something like that instead of a
completely general f(x1,x2); in particular it is equivalent to
(*)     logit(p) = f(x1,x2)
with f(x1,x2) a surface. I'm not aware of anything for R that
includes nonparametric bivariate logistic regression, unless
maybe one of the fitting options in nnet would be appropriate for
your data? tps in funfits can fit a bivariate surface but the only
fitting option is least squares.

 >
 > R-users --
 >
 > I would be interested in tools in R to fit the following gam model:
 >
 > logit(p) = a + f(x1) + f(x2) + f(x1,x2),
 >
 > where f(x1,x2) defines a surface.
 >
 > I have looked into the mgcv library, but it seems only to fit models of
 > the form:
 >
 > logit(p) = a + f(x1) + f(x2)
 >
 > Any ideas?
 >
 > Cheers,
 > Dan
 >



-------------------------------------------------------
Stephen P. Ellner (spe2 at cornell.edu)
Dept. of Ecology and Evolutionary Biology
Corson Hall, Cornell University, Ithaca NY 14853-2701
Phone (607) 254-4221    FAX (607) 255-8088


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._