I was once a consultant for a direct-marketing application wherein I was tasked with building a time series model to predict optimal advertising spend based on a large number of customer contact channels and related variables. I wrote the %ARIMA_SELECT macro to cull out variables that did not contribute significantly to the advertising spend response variable and to present variables that were strongly related to the response variable.
I am sharing the %ARIMA_SELECT macro with the SAS community so that you may benefit from my work.
I presented this work at the WUSS 2013 conference.
Wooow.
The next time I use Forecast Studio (Forecast Server) or Visual Forecasting, I'm definitely going to try this alongside the aforementioned products.
I know SAS R&D worked super hard to select all those (relevant) dynamic regressors and the associated transfer functions correctly (as far as possible).
It seems a bit harsh to me that you can achieve similar results on your own (but I don't want to underestimate you 😉).
Nice work !
Koen
Join us for SAS Innovate 2025, our biggest and most exciting global event of the year, in Orlando, FL, from May 6-9. Sign up by March 14 for just $795.
Data Literacy is for all, even absolute beginners. Jump on board with this free e-learning and boost your career prospects.