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

understanding time series inputs

Super Contributor
Posts: 395

understanding time series inputs

I need some guidance.  I need to forecast retail prices in certain geographical areas, so my dependent variable would be the prices itself, the independent variables could be factors such as 1) # of residents, 2) home prices, 3) unemployment rates 4) GDP .. etc. (the actual variables are just used as examples, not real variables)

My question is, if I want to forecast the prices, then MUSt I also forecast the independent variables if I want to use them in the model.. so eg. forecast the unemployment rate also for the next 12 months?  and question number 2 is using PROC ARIMA, how can I 1) determine which independent variables make sense to use in the model and 2) how and where in the proc arima statement do I insert these independent variables?

Thanks in advance.

SAS Employee
Posts: 8

understanding time series inputs


With respect to a forecast model with inputs - unless you have a situation where the inputs lead the dependent (eg. unemployment goes up but prices doesn't change until 6 months later) you will need to forecast inputs as well.

With respect to the specific ARIMA questions I'd suggest having a look at the documentation: and some of the examples - for instance this:



Ask a Question
Discussion stats
  • 1 reply
  • 2 in conversation