How to estimate a SARIMA?

[lucem_fert]

Hello!

I need to estimate a SARIMA model, but i'm confused about what's the difference among the next models, suppose i have a monthly serie :

1) ls(h) dlog(myserie,1) c ma(1) ar(1) sma(12) sar(12)

2) ls(h) dlog(myserie,1,12) c ma(1) ar(1)

3) ls(h) dlog(myserie,1,12) c ma(1) ar(1) sma(12) sar(12)

I mean if I want to add a sma(12) or sar(12) and I use dlog(myserie,1,12) is it supposed that i can omit "sma" and "sar" terms? And what about if I want to estimate:

3) ls(h) dlog(myserie,1,12) c ma(1) ar(1) sma(12) (without sar(12) )??

Thanks a lot!

[jason_ll]

Interesting question. I was wondering the same too.
In one of my models, I had non seasonally adjusted quarterly data and I was asking myself the same questions.

In the end, I used the dlog(myserie,0,4) specification and removed all sar(4) and sma(4) terms. but I'm interested to hear the experts' responses here.