Hello everybody,
I am a new R user. At current moment I need to do a Principal Components
Analysis with a table who contain mixed variables (categorical and
numerical). There is some function available in R for transform these
variables?
For example: I need transform the categorical variable X
?model.matrix
?contr.treatment
Whether this makes sense for PCA is another matter: princomp() contains a
test to stop this happening automatically with a formula argument. One
problem is the result depends on just how you do the transformation.
On Thu, 24 May 2007, J.Andrés Martínez wrote:
Hi J. Andres,
You are probably better off using the ade4 package, which has two functions
that will do exactly what you want, i.e. a PCA using mixed quantitative and
categorical variables:
## You will need to download ade4 first
library(ade4)
?dudi.hillsmith
?dudi.mix
Regards,
Mark Difford.
Hello again,
I got it with dudi.mix included in ade4 library that Mark told.
Thank you very much to all!
J. Andres
On 5/24/07, Mark Difford [EMAIL PROTECTED] wrote:
Hi J. Andres,
You are probably better off using the ade4 package, which has two
functions
that will do exactly what you