Optimal neural architecture parameters depend on many factors, and there is no simple way to determine what these are short of actually building complete models. Formulas which are often given are merely heursitics and should be taken as, at most, a suggested starting point. Consult the comp.ai.neural-nets (Usenet) FAQ, which can be found at:
ftp://ftp.sas.com/pub/neural/FAQ.html
Specifically, see this section:
ftp://ftp.sas.com/pub/neural/FAQ3.html#A_hu
I will recommend that one shoot for an effective neural network, rather than the optimal one, which will still require some experimentation.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.