Viswanath Devanarayan (DEVANARAYAN_VISWANATH@lilly.com)
Wed, 14 Oct 1998 14:24:38 -0500


I've been using the Splus GAM function to fit logistic additive models for
some data recently.
I usually check to see whether a GAM model (with smoothing spline terms) is
necessary by
comparing the model with a linear term and a smoothing term for a covariate to
the model with
only the linear term for that covariate via Chi-Squared test (with the ANOVA
function, specifying
these two GAM function objects and option test='Chi').

I now have a couple of questions;

1. If it turns out that for one of the covariates a linear term is sufficient,
but for another
covariate a smoothing spline function is necessary, how will I fit this
('semiparametric'?) model using the GAM function (the GAM function doesn't
me any statistic on the covariate with the linear term).

2. What is the best way to deal with categorical covariates in GAM function
where I also have
some continuous covariates with smoothing terms?

Thanks for any help,


This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news