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

Exponential Smoothing Weights and Automatic Forecasting

Accepted Solution Solved
Reply
Frequent Contributor
Posts: 75
Accepted Solution

Exponential Smoothing Weights and Automatic Forecasting

Hello,

In TSFS of ETS when fitting an exponential smoothing model we have 4 options about the values that the smoohting weights can take ([0,1], unrestricted, additive/invertible, intersection of additive/invertible and [0,1]). When TSFS tries to find automatically the best model and the relevant parameters for a time series (fit models automatically option) tests all the four options mentioned above or it assumes that the smoohting weights can take values only between 0 and 1 inclusive?


The same is valid for the Forecast Studio GUI?


Thanks in advance,



Andreas


Accepted Solutions
Solution
‎05-13-2013 02:11 PM
SAS Employee
Posts: 416

Re: Exponential Smoothing Weights and Automatic Forecasting

Hello -

When using the automatic diagnostic option in SAS Forecast Studio for ESM the smoothing weights are restricted between 0.0001 and 0.9999.

In order to apply different weights settings you can:

  • go to "Modeling View" - create a copy of the system generated ESM - and edit your copy. Then you should see the following dialog:ESM.JPG
  • Under Weights you can change the restrictions. Parameters and their restrictions are required to be greater than or equal to –1 and less than or equal to 2.

In order to extend the automatic behavior of SAS Forecast Studio you can add you own ESM repository under "models from an external list" (Model Generation option in Forecast settings). You will need to specify your ESMs using the HPFESMSPEC and HPFSELECT procedures.

Hope this makes sense,

Udo

View solution in original post


All Replies
Solution
‎05-13-2013 02:11 PM
SAS Employee
Posts: 416

Re: Exponential Smoothing Weights and Automatic Forecasting

Hello -

When using the automatic diagnostic option in SAS Forecast Studio for ESM the smoothing weights are restricted between 0.0001 and 0.9999.

In order to apply different weights settings you can:

  • go to "Modeling View" - create a copy of the system generated ESM - and edit your copy. Then you should see the following dialog:ESM.JPG
  • Under Weights you can change the restrictions. Parameters and their restrictions are required to be greater than or equal to –1 and less than or equal to 2.

In order to extend the automatic behavior of SAS Forecast Studio you can add you own ESM repository under "models from an external list" (Model Generation option in Forecast settings). You will need to specify your ESMs using the HPFESMSPEC and HPFSELECT procedures.

Hope this makes sense,

Udo

Frequent Contributor
Posts: 75

Re: Exponential Smoothing Weights and Automatic Forecasting

Hello Udo,

Thanks very much for your answer. Concerning the TSFS the default option when the user specifies an ESM  is zero/one-additive. What about the automatic model fitting facility of TSFS? As i have noticed when ESM are selected the weights are between 0 and 1. Then am i right that the automatic model fitting facility has the default option of [0,1] for the smoohting weights (and hence it is in line with Forecast Studio) and not the zero/one-additive?

Thnaks in advance,

Andreas

SAS Employee
Posts: 416

Re: Exponential Smoothing Weights and Automatic Forecasting

Hello Andreas -

Best to my knowledge TSFS uses zero/one-additive as bounds for the smoothing weights.

Here is how you could verify this yourself:

  • Start the TSFS
  • Select "Fit Models Automatically"
  • Select Options>Model Selection List from the pull-down menu
  • Highlight an ESM - for example "Simple Exponential Smoothing"
  • Right-click the highlighted ESM and select "Edit"
  • This will bring up a dialog which shows you how the model is defined:

ESM.JPG

Hope this helps,

Udo

Frequent Contributor
Posts: 75

Re: Exponential Smoothing Weights and Automatic Forecasting

Hello Udo,

Thank you very much for your answer!

Andreas

☑ This topic is SOLVED.

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

Discussion stats
  • 4 replies
  • 587 views
  • 5 likes
  • 2 in conversation