So I am trying to forecast the price of a certain commodity that is quite volatile. Now I wanted to train a regression feedforward neural network to predict the price of next week.
Just to be clear, I only use lagged values of this commodity's price (including seasonality lags etc.).
Now I read a lot of papers and posts that data scaling/normalizing is very important when you have different features measured in different units (quite logical). However, in my situation all the prices are in euros and there are no external features. I already looked at the first differences of the prices to have values between -1 and 1, apart from computation time the results weren't that different.
Is it still needed to scale/normalize the data?