Neural Network Help

Hello All,

At the moment, I have created a neural network with all the back propagation and so forth. I am now looking at changing the variables such as the weights, learning rate, target and so forth. Are there any good articles out there that suggests what happens, when you change these factors or what they should be set to?

For example I have set my weights to initialize with a random number between -5,5. Is that the correct thing to do?

I understand that the lower you set the learning rate the smaller the increments are each time it iterates, I am guessing I should weigh that up against how much time I want it to take and the complexity of the problem?

Also does the number of hidden layers, reflect on the complexity of the function/problem?

Cheers for any help, or can anyone point me in the right direction?

Thanks CA

----- Edit 1 -----

Cheers, I will have a read and get back to you! :)