Hello
Do we have a differencing functionality in SSM procedure similar to the ones in PROC ARIMA?
For example, in PRCO ARIMA, I can simply define
identify var=sales(1);
to take a simple first difference of the 'sales'.
Then "ESTIMATE statements fit ARMA models to the differenced series. FORECAST statements forecast the differences and automatically sum these differences back to undo the differencing operation specified by the IDENTIFY statement, thus producing the final forecast result."
I'd like to model first difference of my response variable in PROC SSM and have it do the back-transformation for me. How can I accomplish this?
Thanks!
is
Forgot a parenthesis in the SSM syntax:
proc ssm data=test;
trend airline(arima(q=1 sq=1 d=1 sd=1 s=12));
model y=airline;
output out=for;
run;
Unlike ARIMA, the SSM procedure is not interactive and is not designed for interactive Box and Jenkins type model identification process. If you already know the ARIMA model you want to fit (assuming it is of ARIMA(p,d,q)(P,D,Q)s type), you can specify it in the SSM procedure using the TREND statement. In fact, if you are dealing with univariate time series setting, you might find the UCM procedure easier to use. I will provide sample syntax for the monthly airline model for both UCM and SSM procedures:
proc ucm data=test;
model y;
irregular q=1 sq=1 s=12;
deplag lags=(1)(12) phi=1 1 noest; /* specifies non seasonal and seasonal differecing */
estimate;
forecast;
run;
proc ssm data=test;
trend airline(arima(q=1 sq=1 d=1 sd=1 s=12);
model y=airline;
output out=for;
run;
I am providing links for the latest (SAS/ETS 15.1) SSM and UCM docs. Starting with this release you can specify transfer function models in UCM. See the examples in these docs to see additional modeling scenarios.
SSM Doc:
UCM Doc:
Forgot a parenthesis in the SSM syntax:
proc ssm data=test;
trend airline(arima(q=1 sq=1 d=1 sd=1 s=12));
model y=airline;
output out=for;
run;
Available on demand!
Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.
Learn how to run multiple linear regression models with and without interactions, presented by SAS user Alex Chaplin.
Find more tutorials on the SAS Users YouTube channel.