[R] lda

[Prof Brian D Ripley]

On Mon, 6 Dec 1999, Peter Ho wrote:

> This is a problem relating to collinearity of data.

You are failing to give credit for lda, which is from library MASS.

> I have tried using lda on a data matrix. Lda computed the predicted classes, but
> gave a warning that variables were collinear.
> Can anyone comment on this? I am not sure how collinearity of data would affect the
> lda analysis.

It assumes that the data are from multivariate normal distributions with a
common covariance matrix. Such data cannot be collinear. Or, in the
extension of Fisher's scheme, there are many linear combinations maximizing
the quotient of variances.

> What would be the correct procedure to tackle this problem?

Remove the collinearity.


-- 
Brian D. Ripley,                  ripley at 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

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._