When using the ARMA model - Autoregressive Moving Average model using the spikes of ACF and PACF from the correlogram. I
Posted: Wed Jul 06, 2022 5:54 am
When using the ARMA model - Autoregressive Moving Average modelusing the spikes of ACF and PACF from the correlogram.
If I have to use ARIMA to transform the variable to bestationary, which models do we solve for when we want to choose themost suitable model?
Eg: If ARMA (1,3) which was derived from the correlogram whereAR(p) and M(q),
We solve ARMA (1,3), ARMA (1,2) and ARMA(1,1) and compare forthe best suited model.
Therefore, if ARMA(2,3) do we compare models ARMA (2,3), ARMA(2,2) and (2,1) only OR
ARMA (2,3), ARMA (2,2), ARMA(2,1), ARMA (1,3),ARMA (1,2) and (1,1) as there are two AR(p) values.
Or is this process incorrect?
If I have to use ARIMA to transform the variable to bestationary, which models do we solve for when we want to choose themost suitable model?
Eg: If ARMA (1,3) which was derived from the correlogram whereAR(p) and M(q),
We solve ARMA (1,3), ARMA (1,2) and ARMA(1,1) and compare forthe best suited model.
Therefore, if ARMA(2,3) do we compare models ARMA (2,3), ARMA(2,2) and (2,1) only OR
ARMA (2,3), ARMA (2,2), ARMA(2,1), ARMA (1,3),ARMA (1,2) and (1,1) as there are two AR(p) values.
Or is this process incorrect?