R help - Jul 2003 - Non-Linear Principle Component Analysis using Neural Networks

Dear R-help,

I am a new member of the R-help mailing list and would like to know if anyone 
has used R to perform a Non-linear PCA using a multiple layer feed forward 
neural networks? 


Could I modify the function nnet to contain multiple hidden layers? 

Kind Regards,
Paul
_____________________________________________________

Paul Della-Marta
PhD Student                   
Meteorology and Climatology   
Physical Geography            
University of Bern


--------------------------------------------------
This mail sent through IMP at http://mail.unibe.ch

On Fri, 18 Jul 2003, Paul Della-Marta wrote:
> I am a new member of the R-help mailing list and would like to know if
anyone
> has used R to perform a Non-linear PCA using a multiple layer feed forward 
> neural networks? 
> 
> 
> Could I modify the function nnet to contain multiple hidden layers? 
Quite easily: the code is general and there are R-level functions to
specify the net, norm.net and add.net.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595