BookmarkSubscribeRSS Feed
jctran123
Calcite | Level 5

Hi SAS Community, 

 

I am working on a sales forecasting project and I noticed that when I choose Box Cox with a Box Cox parameter of -5, the MAPE improves significantly. Would any one happen to know why? 

 

Thank you in advance, 

Jessica

4 REPLIES 4
ballardw
Super User

Plot transformed data. Look at proc transreg if that isn't what you are currently using.

You'll like find it much smoother and with much less variability.

Possibly over smoothed?

 

Does that MAPE "improve" more as the parameter for the Box-Cox tranform gets further from 0?

jctran123
Calcite | Level 5

Hello @ballardw

 

Yes, the MAPE improves more as the Box Cox parameter becomes a negative number towards -5. It also improves as the parameter moves towards 5, but not as significantly. Please see attached for a screen shot of where I am looking at. Thanks for the suggestion. I have not used proc transreg. Where do I find that? 

 

Jessica 


Capture.GIF
ballardw
Super User

Proc Transreg is in the SAS/STAT package which would be available through the display manager, Enterprise Guide or SAS Studio. I don't know if it is accessible from the Forcast Server as I have never worked with that.

Proc Transreg actually performs the transformations so you can see what happens to your values. It includes plotting functions.

SAS Innovate 2025: Register Now

Registration is now open for SAS Innovate 2025 , our biggest and most exciting global event of the year! Join us in Orlando, FL, May 6-9.
Sign up by Dec. 31 to get the 2024 rate of just $495.
Register now!

Multiple Linear Regression in SAS

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.

Discussion stats
  • 4 replies
  • 1423 views
  • 0 likes
  • 3 in conversation