Hello,
I've recently been using the nnet package to do some basic forecast
predictions. I've found the package to be quite useful and I am
getting some good results. However, I am in the midst of writing a
small paper on the results I am getting and wish to clarify some
things about the nnet package that are not made clear in the
documentation. In particular I would like to know the following:
1) Is it a standard feed forward network trained using gradient
descent (I am assuming this is the case, seems like a no brainer but
just to be sure)?
2) What is the sigmoidal function used for the activation/firing of a
node in the network?
3) What exactly does the output "value" consist of at each iteration?
Is this the value of the Least Mean Square function of the difference
between the output layer and the target values or is it something else?
4) Will this package ever be updated to allow for multiple layers
instead of just one? (just out of curiousity)
I have to present this paper on Friday May 2nd so I would greatly
appreciate a timely response.
Thanks,
-Colin
