Hi All,

I have an input (P), and an output (T) dataset. I wish to capture the
relationship (in unsupervised manner), so that I can subsequently apply that
relationship to another set of dataset. I was trying the following, but I’m
not sure whether there is any better way of doing it. Also, wonder how
should I eventually proceed to actually retrieve the relationship and
implement the same to another set of data. I was mainly trying with two
packages, AMORE and nnet.

net <-
newff(n.neurons=c(1,3,1),learning.rate.global=1e-2,momentum.global=0.5,error.criterium="LMS",Stao=NA,
hidden.layer="tansig",output.layer="purelin",method="ADAPTgdwm")

# Is there any simple way decide and optimize the number of neurons to be
considered?

result <- train(net, P=input, T=target, error.criterium="LMS", report=TRUE,
show.step=100, n.shows=5)

simulated <- sim(result$net, input); joined <- cbind(input,target,simulated)

# above, OK? Not sure about how to proceed next.

Any help/pointer would be appreciated.

Many thanks in advance.

Cheers,

Chintanu

	[[alternative HTML version deleted]]