0
$\begingroup$

How the past error autocorrelation in MA models should be interpreted semantically ? Is it used to capture a relationship that could not be captured by the available regressors ?

$\endgroup$
  • $\begingroup$ Can you provide a little more context for this? It is a little sparse. I'm not sure how well people will be able to answer your question at present. $\endgroup$ – gung Mar 14 '17 at 14:40
  • $\begingroup$ For AR I understand that Y(t) is something * Y(t-1) which make sense as Y(t) is Y(t-1) with more or less a trend or drift or whatever effect. However what can make Y(t) linked to e(t-1) ? $\endgroup$ – user2018454 Mar 15 '17 at 16:25
0
$\begingroup$

The error autocorrelation is expression for the fact that "the error persists for some time until it dissipates".

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.