On Wed, 29 Nov 2000, Thomas Stockton wrote:
> Doesn't the mda library do this? Disr. Analysis using MARS.
No! That's not discriminant analysis in the usual sense, not does it
use the `usual variable selection procedures' for linear or quadratic
discimination. Basically, it does LDA on some predictions from a MARS (or
other) model. Whether that is a useful way to select variables is
debatable, as is whether than LDA is robust enough. (I've published
a couple of alarming examples.)
> Tom
> _
> >-----Original Message-----
> >From: s-news-owner@lists.biostat.wustl.edu
> >[mailto:s-news-owner@lists.biostat.wustl.edu]On Behalf Of Prof Brian D
> >Ripley
> >Sent: Wednesday, November 29, 2000 11:25 AM
> >To: Francois Collin
> >Cc: s-news
> >Subject: Re: [S] Selection of variables in Discr. Analysis
> >
> >
> >On Wed, 29 Nov 2000, Francois Collin wrote:
> >
> >> Hello:
> >> I am running S-plus on Windows NT.
> >> I am working on a classification problem with many feature variables.
> >> Are there functions which implement some of the usual variable selection
> >> procedures?
> >> Thanks in advance for pointers.
> >
> >Not that I know of, and with good reason. They are very dependent on
> >unrealistic assumptions. You would do better to use stepAIC on logistic
> >regression (glm or multinom).
> >
> >--
> >Brian D. Ripley, ripley@stats.ox.ac.uk
> >Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
> >University of Oxford, Tel: +44 1865 272861 (self)
> >1 South Parks Road, +44 1865 272860 (secr)
> >Oxford OX1 3TG, UK Fax: +44 1865 272595
> >
> >---------------------------------------------------------------------
> >This message was distributed by s-news@lists.biostat.wustl.edu. To
> >unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> >the BODY of the message: unsubscribe s-news
> >
> ____________________________________________________________
> Tom Stockton voice: (505) 662-0707 ext 17
> Neptune and Company fax: (505) 662-0500
> 1505 15th St. Suite B stockton@neptuneandco.com
> Los Alamos, NM 87544-3063 LANL mailstop M969
> ___________________ www.neptuneandco.com ____________________
> ---------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu. To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> the BODY of the message: unsubscribe s-news
>
--
Brian D. Ripley, ripley@stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272860 (secr)
Oxford OX1 3TG, UK Fax: +44 1865 272595
|