Which input use to forecast wind speed using MLP

Hello every one,i have a question regarding the neural network (MLP), I would like to know that I could use the maximum and minimum speed as inputs to predict the wind speed. When I've used them, I get a good value from MSE, and vice versa.i forecast the wind speed using(pression ,humidity,temperature as input) and wind speed as target,i have a MLP with 2 hidden layers ,when i add maximal and minimal wind speed in the input i get a goid correlation coefficient(0.998). I really need your answer please. Thanks in advance

[Hiro Yoshino]

Let me clarify your situation:

• Your model: MLP with 2 hidden layers.
• predictors: pressure, humidity and temperature
• response: wind speed

You are thingking of adding the max and the min wind speed as new predictors to improve the performance.

When you added them to the model, you got a good MSE and thus got a good correlation with the ground truth (0.998).

Did I get you correctly so far?

It seems it is working good, what are your worries then?

I think if it works, then it works. Are you worried about an over-fitting?

[Wissal Mass]

Hi Hiro,thank you so much for your answer,yes you get it correctly.

My question is, do I have the right to add the maximum and minimum speeds as inputs to the model, or will I be the one who gives the value of the speed in the inputs, knowing that I want to predict it in the output, as you know that the sum of the maximum and minimum speed divided by two gives the value of the speed I want to expect.

I hope you understand what I mean. Thanks again

[Wissal Mass]

?????

Answer 1

[Mir Sahand]

yes you can