Hi,
there are some dimension reduction methods (somewhat analogous to linear
discriminants) which show groups as separated not only in means, but also
in covariance matrices. This might correspond well to a qda problem.
References are
Young, Marco and Odell, Journal of Statistical Planning and Inference, 17
(1987), 307-319
Fukunaga, Introduction to Statistical Pattern Recognition (2nd ed, p. 455
ff.), Academic Press, 1990.
As I am working recently on more flexible discriminant plots, you may also
consider my technical report on
ftp://ftp.stat.math.ethz.ch/Research-Reports/108.html
where the above mentioned methods are explained as well.
My package fpc on
math.uni-hamburg.de/home/hennig/fixreg/fixreg.html
includes an R implementation of the method of Fukunaga
("Bhattacharyya coordinates"). I have also R code for the method of
Young
et al., but it is not yet "packaged", so you would have to contact me
directly
if you would be interested in a (poorly documented) version.
Best,
Christian
On Thu, 27 Feb 2003, Power, Anne Marie wrote:
> Hi,
>
> I have been using some of the functions in r for classification purposes,
> chiefly lda, qda, knn and nnet.
> My problem is that the only one I can figure out how to represenent
> graphically is lda (using plot.lda). I have tried 'fooling' this
function
> into accepting qda input for plotting but to no avail. I wonder if you
have
> any suggestions?
>
> Thanks alot,
>
> Anne Marie Power
>
> Marine lab.
> Dept. Zooogy & Animal Ecology,
> University College Cork
> Ireland
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> stat.math.ethz.ch/mailman/listinfo/r-help
>
--
***********************************************************************
Christian Hennig
Seminar fuer Statistik, ETH-Zentrum (LEO), CH-8092 Zuerich (currently)
and Fachbereich Mathematik-SPST/ZMS, Universitaet Hamburg
hennig at stat.math.ethz.ch, stat.ethz.ch/~hennig
hennig at math.uni-hamburg.de, math.uni-hamburg.de/home/hennig
#######################################################################
ich empfehle boag.de