[S] Revised: GAM & Supersmoother, S-Plus on Linux on a Mac?

Jean Norris (JCNORRIS@phi.org)
Fri, 04 Dec 1998 09:48:42 -0800


2 modified questions follow:

Has anyone incorporated the Supersmoother into the GAM function call?
GAM currently uses s( ), lo( ), poly( ), not the supersmoother, supsmu( ).
There is a difference of datatype.

I am particularly interested in running 10- to 25-term GAM models on data
with 20,000 (or more) respondents. (The size of the data implies a few
hundred MB of "swap space", a lot of data to move about, and very
large, slow plot files. The results ARE worth it.)

PREVIOUS QUESTION: Has anyone tried running S-Plus on a Mac, with a
G3 or any other motherboard chip, with Linux? Linux is only about $40
for the Mac.

REPLIES include:
1. using a separate Intel motherboard in a Mac box
2. using alternative software ("R") that runs S code on a Mac PPC on top
of LinuxPPC, which results in 'superior memory management'. GAM
routines also would work? (Some of the code may still be in FORTRAN.)

REFRAMED QUESTION: Is anyone out there running S-Plus functions
(including GAM or equally "canned" and complex subroutines)
successfully on Macintosh hardware? Which hardware, which
software? How well is it working?

There are a lot of us who have used, and loved the elegance and ease
of use of, Macs. Can we also have the power and elegance of S-Plus
without a whole computer lab behind us, or becoming sofware and
systems engineers? Thanks for sharing your experiences, or wise
thoughts, on the topics.

Jean Norris
Research Scientist
Public Health Institute
Berkeley, CA
(510) 649-1987
email: JCNorris@phi.org
alternative email: jeann@stat.berkeley.edu

-----------------------------------------------------------------------
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