I would like to request your assistance in interpreting this model in Time series. $$ (1 − B) (1 − B_{4})(1 − 0.43B_4)X_t = (1 + 0.22B)(1 + 0.88B_4)W_t. $$
$\begingroup$
$\endgroup$
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
The seasonal ma polynomial (coeff =.88) is effectively cancelling the seasonal difference . I suggest that you simplify your model by eliminating the seasonal difference irrespective of the poor guidance by your software's overfitting . Make sure that the model errors don't suggest pulses and/or level shifts or other gaussian requirements otherwise you have effectively mis-modelled.
$\begingroup$
$\endgroup$
Time series models usually do not have interpretation the same way that cross section econometrics has.
What you can interpret (assuming you have a series with quarterly frequency) is that the model is a SARIMA.