Hello,
I'm trying to do some correspondence analyis, with R, of course (by
correspondence analysis, I'm refering to JP Benz?cri's methods, in case
there might be some other thing with a similar name)
I've found a couple of tools refering to C.A in the existing packages :
ca() in package multiv and corresp()/mca() in MASS. MASS tools look more
easy to use (it is supposed to put the data under disjunctive form by
itself, and some high level plotting functions are provided). On the
other hand, ca's results look closer to what I've seen in data analysis
textbooks.
What is worse, I've tried ca() and corresp() on a sample dataset, and
could not get the same results (or rather, I was not able to find a link
between the two result sets) :o(
First question: which one of these packages do you advise me to use?
Second question: if I use MASS, how can I get plots of the different
factorial plans. The doc says something I find very weird: "nf: The
number of factors to be computed. Note that although 1 is the most
usual, one school of thought takes the first two singular vectors for a
sort of biplot." All textbooks I've read on correspondence analysis
insisted on the fact that this method gives simetrical roles to the 2
analized factors (and therefore a plotting both factors on the same
graph was natural), and that it was necessary to take the first n
factorial axis into account when interpreting the results.
Any insight on these matters is welcome.
Alexandre
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
