electricity price forecasting using neural network

I'm doing my final year project on short term electricity price forecasting using neural network. i need to varying the number of hidden neurons, learning rate and momentum rate. can anyone help me how to write the coding on varying all those parameters? thank you for your help :)

Answer 1

[Greg Heath]

Start with the default code. Run 10 times with different initial weights.

If successful, repeat with smaller number of hidden nodes

If unsuccessful, repeat with larger number of hidden nodes

Find the smallest number of hidden nodes that yields acceptable results.

Fix the number of hidden nodes and vary the other two parameters over a 2-D grid.

Hope this helps.

Thank you for formally accepting my answer

Greg