This is a modest framework for dealing with neural network. This is not a
big project, I have done this just to learn a bit about neural networks.

I have got some papers and two books about the subject from a friend of mine:

o. John Hertz, Anders Krogh, Richard G. Palmer, "Introduction to the
Theory of Neural Computation", Santa Fe Institute studies in the science of
complexity, lecture notes, computation and neural system series, 1991.
Addison-Wesley.

o. Simon Haykin, "Neural Networks: A Comprehensive Foundation", 1994,
Macmillan College Publishing Company.

There are not too many comments, sorry. Anyway, I beleave the class names,
instance variables and protocol will give you a good idea of how this works.
BTW, I have put a little example at the end of the method
BackPropagationTrainer class|xor; just cut-and-paste it in a workspace and
try it out. (it converges almost always, but not always... for sure, there
are a bug somewhere...).

Luciano.- (lnotarfr@dc.uba.ar)