BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
omerzeybek
Obsidian | Level 7

Hi i am trying to estimate next best offer for my every customer. Bur my customers have some demographics segments and i want predictions to be  made in all segments seperately on its own. 

 

For regression analysis i was able to estimate n regressions for n different  segment, when i coded segment variable. Bu i didnt able to carry same method for link analysis. 

 

do you know any short way for doing  this. 

1 ACCEPTED SOLUTION

Accepted Solutions
rayIII
SAS Employee

OK, so I spent some time trying to get a simple example to work in EM and kept running across errors until I saw this explanation in the documentation: 

 

"You cannot perform group processing on Transaction data sets in SAS Enterprise Miner."

 

So if you are using transaction data, that's the issue.

 

Do you have a lot of segments? f you are using transaction data you'll probably have to build a separate link analysis subflow for each segment. 

 

 

View solution in original post

9 REPLIES 9
rayIII
SAS Employee

How did you build your segmented regression models? Using the group processing nodes? 

 

Ray

omerzeybek
Obsidian | Level 7

As  do i remember, when i kept my segment identifier variable in variables in segment role, i was receiving different regressions for every different segment... 

rayIII
SAS Employee

Hi.

 

It looks to me like the Regression and HP Regression nodes ignore segment variables. 

 

Is it possible you were actually building stratified models using the Start Groups and End Groups nodes as in this tip? That's how I would normally go about it. 

 

(Possibly this was supported in an early version or EM. I'm using 14.1)

 

Ray

omerzeybek
Obsidian | Level 7

unfortunately when  i add a start group node just before my link analysis node, i am unable state my segment variable as grouping rule. How can i do this? 

WendyCzika
SAS Employee

Make sure your segment variable has Level=Nominal (In Input Data node, or a Metadata node).

omerzeybek
Obsidian | Level 7

Thank you for your comment 

But it is in Nominal Level 

rayIII
SAS Employee

OK, so I spent some time trying to get a simple example to work in EM and kept running across errors until I saw this explanation in the documentation: 

 

"You cannot perform group processing on Transaction data sets in SAS Enterprise Miner."

 

So if you are using transaction data, that's the issue.

 

Do you have a lot of segments? f you are using transaction data you'll probably have to build a separate link analysis subflow for each segment. 

 

 

omerzeybek
Obsidian | Level 7

Correct Raylll

 

I saw exactly same mesagge in EM. But as you know i must use  transactional data to perform a next best offer model. As you noted i started building seperate data files for my 12 different  segments. 

 

Thanks a lot 

rayIII
SAS Employee

Good luck. Sorry I didn't see that sooner. 

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

Register now!

How to choose a machine learning algorithm

Use this tutorial as a handy guide to weigh the pros and cons of these commonly used machine learning algorithms.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 9 replies
  • 1573 views
  • 1 like
  • 3 in conversation