About number of weights and hidden layer in neural network MATLAB 2012b

Dear Prof Greg,

I found the followings from answers in MATLAB replied by you--

""Find a good value for the number of hidden nodes, H, by trial and error. Create numH*Ntrials nets in a double loop: An outer loop H = Hmin:dH:Hmax over number of hidden nodes and an inner loop i = 1: Ntrials over number of random trn/val/tst data divisions and random weight initialization trials for each value of H.

I typically use numH ~ 10, Ntrials ~ 10 and tabulate results in Ntrials X numH matrices.

Many examples can be obtained by searching the NEWSGROUP and ANSWERS using the keywords""

Greg Ntrials Nw""

My question: 1) Is there any books/links /websites refering the above mathematical equations of Nw, Ntrnq etc ? Or this is an emperical formulae(i.e just by experience) Please pardon if it is an inane query.thanks a lot.

Answer 1

[Greg Heath]

Most of what I post is home grown from my 30 years of experience with neural nets. I have posted in comp.ai.neural-nets as well as MATLAB NEWSGROUP and ANSWERS. I try to keep my notation as consistent as possible. However, changes do happen.

So, if you want a better or just different look/explanation of something in one of my postings. Try searching and listing posts from the present backwards. I have used the following email addresses

 [e-mail]

 [e-mail]

 [e-mail]

 [e-mail]

If you still have questions after a reasonable search, just post on NEWSGROUP or ANSWERS. I don't use comp.ai.neural-nets much anymore.

Greg