[R] Canonical Variance Analysis by any other name?
Prof Brian Ripley
ripley at stats.ox.ac.uk
Thu Jan 19 09:35:10 CET 2006
Canonical variates or canonical variance?
Canonical variates in the sense of your quote is the same thing as
multi-group LDA (the part of LDA not due to Fisher)
Another sense (e.g. http://en.wikipedia.org/wiki/Canonical_analysis)
is canonical correlation analysis (function cancor). And indeed, LDA can
be derived from CCA: see my PRNN book.
On Thu, 19 Jan 2006, Patrick Connolly wrote:
> I've been asked about "Canonical Variance Analysis" (CVA). I don't
> see any reference to it searching the R site. Does it go by other
> Genstat describes it thus:
> Canonical variates analysis operates on a within-group sums of squares
> and products matrix, calculated from a set of variates and factor that
> specifies the grouping of units. It finds linear combinations of the
> variates that maximize the ratio of between-group to within-group
> variation, thereby giving functions that can be used to discriminate
> between the groups.
> It's probably not particularly difficult to do, so I suspect someone
> has a package for doing it. What other name might I search for?
> Patrick Connolly
> Mt Albert
> New Zealand
> Ph: +64-9 815 4200 x 7188
> I have the world`s largest collection of seashells. I keep it on all
> the beaches of the world ... Perhaps you`ve seen it. ---Steven Wright
> R-help at stat.math.ethz.ch mailing list
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
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 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595
More information about the R-help