[R] CPCA?

[Ben Bolker]

Jari Oksanen wrote:
> 
> Daniel Berner <daniel.berner <at> mail.mcgill.ca> writes:
> 
>> 
>> It would be great to know if and where an R code for Common Principal
> Component Analysis is available.
> 
> The only code I know is in IMSL Fortran library, and that is proprietary,
> and
> that is expensive. After a quick look, it really appears that IMSL C
> library
> does not have the code, but you got to go to Fortran. Look for function
> DKPRIN
> (KPRIN). That code cannot be ported to R, naturally.
> 
> Another issue is that the code obviously was written by Bernhard Flury,
> and he
> also has FORTRAN listing in his book "Commaon Principal Components and
> Related
> Multivariate Models" (J. Wiley & Sons, 1988). I don't know of the
> licensing
> status of that code. I didn't find any explicit licensing information in
> the
> book (but may have overlooked something), apart from the usual book
> copyright
> that forbids everything. I assume it is the same code that is the base of
> DKPRIN
> in the IMSL FORTRAN library. Would mean a lot of typing at minimum, and
> for any
> sensible code, editing the code to use LAPACK where ever possible.
> 
> Further, Flury descirbes the algorithm in his book, and it could be
> implemented
> independently in R. I think nobody has done that (yet).
> 
> cheers, jari oksanen
> 
> 

I have a package that wraps some code Patrick Phillips wrote -- haven't put
it on CRAN (yet) because of issues with including executable files in
packages ...
e-mail me if you want further info.

  Ben Bolker

-- 
View this message in context: http://www.nabble.com/CPCA--tf4873113.html#a13948778
Sent from the R help mailing list archive at Nabble.com.