This question may have been asked in various formats but I'm struggling understanding the reasons why, so here it is:
Why is the forecast from my ARIMA model a simple straight line? The strange thing, is that for the historical data, it is an almost perfect fit, but as soon as I step outside that historical range, it just produces a straight line. I would understand if it was a straight line all along, but not if it's only after being out of the historical data.
I'm using an ARIMA(3,1,3) model on a commodity which had a linear trend. I included a graph of the situation to make it more clear
. The code is very short
import pandas as pd
from statsmodels.tsa.arima.model import ARIMA
series = pd.DataFrame(some_data)
model = ARIMA(series, order=(3,1,3))
result = model.fit()
predictions = result.predict(start=len(WTI_filled)-150, end=len(WTI_filled)+50)
series.plot()
predictions.plot()
