Page 1 of 1

ARMA estimation

Posted: Thu Nov 18, 2010 9:08 am
by ranpan
Hi !

I am using version 7. I need to estimate an ARMA(p,q) model with linear trend and seasonal dummies. I am not quite clear about the form it will take. Could you give me some indications please ?

Re: ARMA estimation

Posted: Sat Nov 20, 2010 4:49 am
by ranpan
In continuation, what is the significance of seasonals SAR(1) - SAR(4). Does it mean that SAR(4) accounts for seasonality in qtr 4 ?

I am hoping to get some pointers on my earlier query too. Thanks in advance.

Re: ARMA estimation

Posted: Sat Nov 20, 2010 6:14 am
by ranpan
Suppose I want to fit an ARMA(p,q) to a variable with linear trend and seasonality : X c ar(1) ma(1) SAR(4) SMA(4) ?

I have fitted many more with different AR and MA terms. The best model would depend on the analysis of the estimation outputs?

Thanks for your time

Re: ARMA estimation

Posted: Sun Nov 21, 2010 2:32 pm
by ranpan
Hi!

Is this a valid estimation equation - X c @trend ar(1) ma(1) SAR(4) SMA(4) ?

Yes, I am still learning, so pardon the trivial nature of my queries. I hope someone is good enough to have a look at these queries of mine in a positive way.

Thanks

Re: ARMA estimation

Posted: Sun Nov 21, 2010 4:07 pm
by startz
Sure, it's a valid equation. It may or may not be a good model. That depends in part on the nature of the data and what you're going to do with the estimate after you have it. But there's nothing inherently invalid in what you're doing.