Large values of MSE and RMSE

Dear Greg Hope you are doing fine and well. I have read your kind comments on “large values of mse and/or rmse”. But, I cannot implement your mentioned steps in matlabnews. I am working on time-series solar energy forecasting. I have the target data, and when I train the network, it gives me a very high value of mse and rmse. So what should I do before feeding the target data to the neural network and training it? I think that I should manipulate the target data, in such a way that it should ultimately get the following NMSEgoal: NMSEgoal <= 0.01 Looking forward to hearing from you. Warmly,

Answer 1

[Greg Heath]

I always plot the auto and cross correlation functions before determining which range of delays to try during my search of the minimum number of hidden nodes.

Hope this helps.

Thank you for formally accepting my answer

Greg

[Hisham Hussein]

Thanks a lot Greg Would you kindly write me the steps for plotting the auto and cross correlation functions and the steps also for searching for the minimum number of hidden neurons. I am really finding difficulties in searching of these steps through your posts and answers. Plz excuse me for any inconvenience cause by me.

[Greg Heath]

Sorry, I would need to take just as much time as you.

The real difference is that I would not learn anything from it.

[Hisham Hussein]

Never mind and thanks a lot for your kind efforts. I will try to do it by myself.