Dear R users, 

I m quite a novice in using R for factor analysis and I would need some help to choose 
the right function. 
I have a contingency table and I would like to perform a  Correspondence analysis on 
this table, followed by a hirarchical clustering of my variables projected in on the 
first principal components.

Here are my question : 

- what is the more appropriate function to do so ... I already tried by using the 
function 'corresp' in the MASS package
- it seems to work ... but how is it possible to get all the information concerning 
the non-centered PCA used by corresp (eigen-values, inertia, scree plot, square cos, 
....)

In a second step I would like to use hirarchical clustering from the results of the 
correspondence analysis ... 

If "Table" is my contingency table (i.e., the number of individuals seen in each case) 
... I tried to implment it as follow (for the 2 first components for instance, but I 
would be interesting at looking to the other components ...I did not manage to get the 
eigen values !) :

A <- corresp(Table, nf = 2)
biplot(A)
hc <- hclust(dist(A$cscore), "ward")
plot(hc)

Is that ok ??? ... here it is exemplae using Ward method for clustering ... 
I would also be interested in using method of mutual neighbors  to identify the 
clusters ... is it possible using hclust ?

Any help would be really appreciated,

Best regards ... 

Christophe

***************************
Christophe Grova, PhD 
PostDoc - EEG department
Montreal Neurological Institute, McGill University
3801 University Street, Montreal, Quebec, Canada, H3A 2B4
email : [EMAIL PROTECTED]
tel : (514) 398 2184
fax : (514) 398 8106
web: http://idm.univ-rennes1.fr/users/grova
*************************** 

        [[alternative HTML version deleted]]

______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

Reply via email to