Can anyone help me with a detailed workflow for using NN toolbox for predicting day-ahead solar irradiance from historical hourly time series?

I am presently working on my undergraduate final year project and part of it entails a day-ahead 24h prediction of solar irradiance from historical hourly time series. I have never applied the neural network toolbox for anything of this sort before, so I am really finding it difficult using it to solve my problem. I have gone through chapters 1-3 of the user guide, yet I still don't know exactly how to start using it to address my problem.

 I will greatly appreciate help from anyone who can be patient enough to help me out.

Answer 1

[Greg Heath]

It depends on whether you want to use output feedback or not. Fot no feedback

 help timedelaynet  %prediction depends on current and past input
 doc timedelaynet

 help narxnet        %prediction depends on past output in addition to current and past input
 doc narxnet

Search the NEWSGROUP and ANSWERS using a subset of the following search words

 greg narxnet prediction
 greg narxnet forecast

Hope this helps.

Thank you for formally accepting my answer

Greg