Forecasting using SAS Forecast Server, SAS/ETS, and more

Seasonal Factors

Accepted Solution Solved
Reply
Occasional Contributor
Posts: 5
Accepted Solution

Seasonal Factors

What have you found to be the best way to determine seasonal patterns in data?


Accepted Solutions
Solution
‎07-08-2014 10:53 AM
Super Contributor
Posts: 340

Re: Seasonal Factors

Posted in reply to gwcanalytics

I think one statistally right way would be a seasonal Unit-Root-Test:

Proc Arima Data=<Data_Set>;

  Identify Var=<Variable> Stationarity=(ADF=(1) DLag=12); /* Augmented Dickey-Fuller-Test with 1 autoregressive term; monthly data */

  Identify Var=<Variable> (1) Stationarity=(ADF=(1) DLag=12); /* -"- in 1st differences */

Run;

There are other tests as well (and simpler methods). For example if you have quaterly data and carry out a trend regression you might check the dummy-coefficients for statistical significance.

View solution in original post


All Replies
Trusted Advisor
Posts: 1,228

Re: Seasonal Factors

Posted in reply to gwcanalytics

I think we can't generalize which is the best way but as an initial point, data on a graph can give information about seasonal/cyclic patterns.

Solution
‎07-08-2014 10:53 AM
Super Contributor
Posts: 340

Re: Seasonal Factors

Posted in reply to gwcanalytics

I think one statistally right way would be a seasonal Unit-Root-Test:

Proc Arima Data=<Data_Set>;

  Identify Var=<Variable> Stationarity=(ADF=(1) DLag=12); /* Augmented Dickey-Fuller-Test with 1 autoregressive term; monthly data */

  Identify Var=<Variable> (1) Stationarity=(ADF=(1) DLag=12); /* -"- in 1st differences */

Run;

There are other tests as well (and simpler methods). For example if you have quaterly data and carry out a trend regression you might check the dummy-coefficients for statistical significance.

Super Contributor
Posts: 340

Re: Seasonal Factors

Posted in reply to user24feb

P.S.: I guess you are aware that your question is a tricky one. If you ask your question this way, it is almost certain that the respondent puts his/her statistical reputation on the line. :-)

🔒 This topic is solved and locked.

Need further help from the community? Please ask a new question.

Discussion stats
  • 3 replies
  • 535 views
  • 0 likes
  • 3 in conversation