1
$\begingroup$

The first uploaded image presents the ACF and PACF after the series differentiation. I observed the pattern at lags 7, 14, 21 etc. and recognized it as the seasonal pattern, having the original data that was "daily"

The second picture presents the ACF and PACF after seasonal adjustment. Now i need to choose parameters p and q for ARIMA model. (I(d) is already taken care of as i did one differentiation thus d = 1).

The first uploaded image presents the ACF and PACF after the series differentiation. I observed the pattern at lags 7, 14, 21 etc. and recognized it as the seasonal pattern, having the original data that was "daily".

The second picture presents the ACF and PACF after seasonal adjustment. Now i need to choose parameters p and q for ARIMA model. (I(d) is already taken care of as i did one differentiation thus d = 1)

So, generally i am looking for help with such a problem. I described it in description of the uploaded photos.

$\endgroup$
  • $\begingroup$ The first picture represents the ACF and PACF for differened series. Noticed the seasonality at lag 7, 14, 21 etc. as the data is "daily". The second picture presents the ACF and PACF after seasonal adjustment. Now i need to choose parameters p and q for ARIMA model. (I(d) is already taken care of as i did one differentiation thus d = 1). $\endgroup$ – Karol Śmigielski Jan 3 '17 at 1:46
  • 1
    $\begingroup$ I barely follow your question, & I can't read you pictures. I suspect if I could understand this, it would be too broad to be answerable. Please try to add sufficient context to make this understandable, & try to make it narrower & more concrete than "I am looking for help". $\endgroup$ – gung - Reinstate Monica Jan 3 '17 at 1:52
  • $\begingroup$ Now it's how it should be. I got a bit lost (late hours kicked in). Now the post is edited and i hope my question is relatively narrow (I'm new into this science so cannot be sure). $\endgroup$ – Karol Śmigielski Jan 3 '17 at 2:09
  • $\begingroup$ Please add new information as edits to the post, not as comments! $\endgroup$ – kjetil b halvorsen Jan 3 '17 at 2:09
  • $\begingroup$ Yes, yes. My bad. Already edited. $\endgroup$ – Karol Śmigielski Jan 3 '17 at 2:10
0
$\begingroup$

Since there appears to be one significant acf value at period 7 in the ACF function this might suggest the need for an MA seasonal component. You should know that more often than not when dealing with daily data the MORE correct approach is to consider incorporating daily effects like specific days-of-the-week , days-of-the-month , week/month of the year , lead and lag effects of holidays etc. Try searching for DAILY in SE.

$\endgroup$

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.