Mapping R ARIMA parameters to coefficient symbols

Asked Mar 07 '18 at 11:51

Active Mar 07 '18 at 17:15

Viewed 61 times

I'm trying to map the information R prints for an ARIMA model to the coefficients in the formulas that I'm familiar with. Here's what I have so far

where an AR(p) model is defined as

$X_t = c + \sum_{i=1}^p \varphi_i X_{t-i}+ \varepsilon_t$

and an MA(q) model is given by

$X_t = \mu + \varepsilon_t + \sum_{i=1}^q \theta_i \varepsilon_{t-i}$

Note: I have read Definitions of coefficients from Arima {forecast}, but it doesn't discuss the "intercept" and "mean" parameters.

edited Mar 07 '18 at 17:15

Richard Hardy

asked Mar 07 '18 at 11:51

mhwombat

1

See Issue 3 by Stoffer. – Richard Hardy Mar 07 '18 at 12:38
@RichardHardy, if you want to convert that to an answer, I'll accept it. – mhwombat Mar 07 '18 at 16:58
I believe the question is a duplicate as these issues have been brought up a few times before. I will see if I can find the relevant thread(s). If not, I can probably write it up as you suggest. – Richard Hardy Mar 07 '18 at 17:02
What about this? – Richard Hardy Mar 07 '18 at 17:06
@RichardHardy Yes, mine is a duplicate of that question. – mhwombat Mar 07 '18 at 17:07
Let's mark it as such then. – Richard Hardy Mar 07 '18 at 17:14
1

Note one important aspect: there are two prominent conventions for MA models, which differ in the signs of the coefficients, so don't be surprised if your MA coefficients have their signs flipped. – Stephan Kolassa Mar 08 '18 at 15:57

0 Answers0