BookmarkSubscribeRSS Feed
PaalNavestad
Pyrite | Level 9

Hi we are working on a workflow where we have sporadic measurements. The data is not stationary with trend changes at infrequent intervals. There are no natural cycles in the data. We have found that smoothing the data and checking the residuals for outliers seem to work well.

 

To make the workflow as efficent as possible we would like to run the whole flow in CAS. 

 

However we have not found any CAS methods for LOESS or Bspline or Mars regressions (adaptive reg).

 

Can anybody please guide us in the right direction ?

3 REPLIES 3
StatDave
SAS Super FREQ

In SAS Viya, you could either use PROC GAMMOD (or GAMSELECT if model selection methods are needed), or if you prefer to use CAS actions, you can use the related actions in the Generalized Additive Model action set. These allow fitting a model using thin-plate regression splines. See the SAS Viya documentation for details.

Rick_SAS
SAS Super FREQ

In general, if you are looking for computations on the CAS server that are similar to what you would find in SAS/STAT procedures, look at the 

Similar documentation exists for CAS-enabled econometrics/time series procedures and the econometric actions.

As statdave_sas says, I'd recommend PROC GAMMOD.

Ready to join fellow brilliant minds for the SAS Hackathon?

Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. Registration is now open through August 30th. Visit the SAS Hackathon homepage.

Register today!
What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 3 replies
  • 387 views
  • 2 likes
  • 3 in conversation