R help - Nov 2008 - Input correlation matrix directly to princomp, prcomp

Hello fellow Rers,

 

I have a no-doubt simple question which is turning into a headache so
would be grateful for any help.  

 

I want to do a principal components analysis directly on a correlation
matrix object rather than inputting the raw data (and specifying cor TRUE or the
like).  The reason behind this is I need to use polychoric
correlation coefficients calculated with John Fox's hetcor function. Is
there a way to do this with princomp or prcomp, or any other principal
components function in other R packages?

 

Again, very grateful for any help.

 

David Evans. 

 


	[[alternative HTML version deleted]]

On Mon, 3 Nov 2008, EVANS David-William wrote:
> Hello fellow Rers,
>
>
>
> I have a no-doubt simple question which is turning into a headache so
> would be grateful for any help.
>
>
>
> I want to do a principal components analysis directly on a correlation
> matrix object rather than inputting the raw data (and specifying cor >
TRUE or the like).  The reason behind this is I need to use polychoric
> correlation coefficients calculated with John Fox's hetcor function. Is
> there a way to do this with princomp or prcomp, or any other principal
> components function in other R packages?
See ?princomp, and

   covmat: a covariance matrix, or a covariance list as returned by
           'cov.wt' (and 'cov.mve' or 'cov.mcd' from
package 'MASS'). If
           supplied, this is used rather than the covariance matrix of
           'x'.

A correlation matrix *is* a covariance matrix (of scaled data, and scaling 
does matter for PCA, of course).

That this is possible is the main advantage of princomp over prcomp.
Compare
> princomp(covmat=cor(USArrests))
Call:
princomp(covmat = cor(USArrests))

Standard deviations:
    Comp.1    Comp.2    Comp.3    Comp.4
1.5748783 0.9948694 0.5971291 0.4164494

  4  variables and  NA observations.

with the help-page examples
>
>
>
> Again, very grateful for any help.
>
>
>
> David Evans.
>
>
>
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
-- 
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