I am studying a time series using auto_arima to determine the best parameters for my SARIMAX model. The code is the following:
auto_arima(df['Sample'].loc[Date_from:Date_until],start_p=0,start_q=0,max_p=2,max_q=2,
seasonal=True,exog = df[['Sample_exog']].loc[Date_from:Date_until],
trace=True,enforce_stationarity=False,enforce_invertibility=False, m=7)
Auto_arima is trying to find the solution which reduces the AIC, but I do not understand why is not taking into account the solutions with the term "intercept". What does it mean?
I am also running a rmse test and the best models are the ones with the "intercept" word in auto_arima. Should I reject these solutions?
Thank you