model inputs

[mmontrond]

Why all of the work done in detrending and making the model stationary is not used. The original datasets is what is used as inputs to all models?
After you performed Combining Detrending and Differencing, you commented "This is the transformation we will use moving forward with our analysis". However, I don't see that used anywhere Would you please explain?

[superjeje]

Hi, as written above, we cannot see where your transformed data are used. An example with these transformed data using an algorithm non dedicated to TS, could be interesting (RF Regressor for instance) : how do you retrend and undifferentiate once you get the predictions with those data?