Hi All,
I have an ARIMA model that was generated by Forecast Studio. The model has an input series (independent variable) used to forecast the dependent variable. As I have not provided any future values for the input series, Forecast Studio must be generating a forecast for the input series which it then feeds into the model for the dependent variable. I am trying to find the specific model formulation used to generate a forecast for the input series but cannot figure out how to see this in Forecast Studio.
Any help would be appreciated.
To be a bit more specific, by default Forecast Studio uses one of the following smoothing models to extend the future input variable values:
• simple
• double
• linear
• damped trend
• seasonal
• Winters method (additive and multiplicative)
The model with the smallest in-sample MAPE is used to forecast the future values of the input.
You can use PROC HPF method = BEST to reproduce the Forecast Studio result if you have the Forecast Server license. If you only have the license for Forecast Studio Desktop version, you will have to run PROC ESM for all the models listed above and pick the model with the smallest in-sample MAPE.
I would suggest feeding the Forecast Studio with future input values (and set the lead = 0) if you do have them available. Often time it is inappropriate to forecast future input variable values if they are not stochastic. For example, promotions, events, etc. Having Forecast Studio to generate forecasts for non-stochastic input variables could distort the shape of the forecasts into the future.
use PROC ESM to that input variable.
To be a bit more specific, by default Forecast Studio uses one of the following smoothing models to extend the future input variable values:
• simple
• double
• linear
• damped trend
• seasonal
• Winters method (additive and multiplicative)
The model with the smallest in-sample MAPE is used to forecast the future values of the input.
You can use PROC HPF method = BEST to reproduce the Forecast Studio result if you have the Forecast Server license. If you only have the license for Forecast Studio Desktop version, you will have to run PROC ESM for all the models listed above and pick the model with the smallest in-sample MAPE.
I would suggest feeding the Forecast Studio with future input values (and set the lead = 0) if you do have them available. Often time it is inappropriate to forecast future input variable values if they are not stochastic. For example, promotions, events, etc. Having Forecast Studio to generate forecasts for non-stochastic input variables could distort the shape of the forecasts into the future.
Registration is open! SAS is returning to Vegas for an AI and analytics experience like no other! Whether you're an executive, manager, end user or SAS partner, SAS Innovate is designed for everyone on your team. Register for just $495 by 12/31/2023.
If you are interested in speaking, there is still time to submit a session idea. More details are posted on the website.
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.