Page 1 of 1
Representation view with AR(1)
Posted: Tue Dec 02, 2014 11:45 am
by startz
It might be good if the representation view after
looked like
X = C(1) + [AR(1)=C(2)]*RESID(-1)
rather than like
X = C(1) + [AR(1)=C(2)]
Re: Representation view with AR(1)
Posted: Wed Dec 03, 2014 11:01 am
by EViews Glenn
Interesting idea, and I can see why it might be desirable, but then we'd have to have lots of lags of RESID running around for higher order and seasonal polynomial models. One more thing, changing the representation is not innocuous as it's used by the forecast and model engine, which would have to be rewritten to accommodate the new format. And that's not even talking about MAs.
But perhaps there's another place where we could display the expanded polynomial forms or perhaps forms using the L operator.
Re: Representation view with AR(1)
Posted: Wed Dec 03, 2014 11:04 am
by startz
I was really only thinking about the display issue. I've had some intro students who read the current representation and misunderstood what it implied for fitted values.
Re: Representation view with AR(1)
Posted: Fri Dec 05, 2014 11:53 am
by EViews Glenn
So maybe an addition to the ARMA structure view...