This directory contains code and documentation for an R package based
on `nnet' from Venables and Ripley, `Modern Applied Statistics with
S-Plus' (2nd edition).

The S original can be downloaded from
  http://www.stats.ox.ac.uk/pub/MASS2/Unix.shtml
and is Copyright (C) 1992-8 by by W. N. Venables and B. D. Ripley and
released under the GPL (see the LICENCE file in the original for further
details).  

The original README file is appended below.

The directory `man' contains R documentation sources converted via Sd2Rd
from the S documentation sources.

See `PORTING' for information on the changes in the R version.

KH <Kurt.Hornik@ci.tuwien.ac.at> 1998/02/22.


*** Original README ****************************************************

nnet library from Venables & Ripley (2nd edition)
=================================================

Software for feed-forward neural networks with a single hidden layer.

nnet		fit neural network using variable metric optimizer in C
predict.nnet	predict future observations.
print.nnet
summary.nnet
nnet.Hess	evaluate Hessian

which.is.max	select maximum of vector, breaking ties at random
class.ind	generates indicator matrix for factor

software for multinomial log-linear models

multinom
predict.multinom
print.multinom
summary.multinom
add1.multinom
drop1.multinom
coef.multinom
vcov.multinom
max.col		find largest column for each row of a matrix,
		breaking ties at random
