R help - Aug 2009 - principal component analysis for class variables

Dear Forum,

I have a class variable 1 (populations A-E), and two other class variables,
variable 2 and variable 3. What I want is to see if the combination of var 2
and var 3, will give me a pattern that allows to distinguish populations. 
I found several packages like ade4, with pcaiv function and factoMineR. but
there are not working. Using the ade4 package, when I try to build the pca:
 pca1 <- dudi.pca(D, scan = FALSE, nf = 2)
Error in v * row.w : non-numeric argument to binary operator

Does someone has suggestions? 
Thanks,
Andy
-- 
View this message in context:
http://www.nabble.com/principal-component-analysis-for-class-variables-tp24792474p24792474.html
Sent from the R help mailing list archive at Nabble.com.

The PCA doesn't work with class variables so the error is normal. You 
should try to work with a discriminant factorial analysis (see 
discrimin.coa in ade4).

Alain



andreiabb wrote:
> Dear Forum,
>
> I have a class variable 1 (populations A-E), and two other class variables,
> variable 2 and variable 3. What I want is to see if the combination of var
2
> and var 3, will give me a pattern that allows to distinguish populations. 
> I found several packages like ade4, with pcaiv function and factoMineR. but
> there are not working. Using the ade4 package, when I try to build the pca:
>  pca1 <- dudi.pca(D, scan = FALSE, nf = 2)
> Error in v * row.w : non-numeric argument to binary operator
>
> Does someone has suggestions? 
> Thanks,
> Andy
>   
-- 
Alain Guillet
Statistician and Computer Scientist

SMCS - Institut de statistique - Universit? catholique de Louvain
Bureau d.316
Voie du Roman Pays, 20
B-1348 Louvain-la-Neuve
Belgium

tel: +32 10 47 30 50