topic Re: By-group processing in E-Miner in SAS Data Science

By-group processing in E-Miner

kdp — Mon, 29 Jun 2009 21:24:07 GMT

Hi,

I have E-Miner version 5.2 and I am trying to do by-group processing similar to E-Guide.

Is it possible to even do this E-Miner? If so, what roles must the variables be assigned to do this?

Thanks,
kdp

Re: By-group processing in E-Miner

Sascha_Schubert — Thu, 09 Jul 2009 07:45:41 GMT

Hi kdp,

you can use the group-processing nodes in SAS EM to do this. You use a Start Group and an End-Group Node and put the nodes that need to be included in the loop in between the start and end node. Set the property "Mode" in the Start group node to "Stratify" and then open the Variables window. In there set the property "Grouping Role" of your index variable to "Stratification". For example, a variable like gender with 2 classes would then create 2 models automatically.

For more information refer to the SAS EM online help in the section Node Reference --> Utility --> Start Group Node.

Thanks
miner40

Re: By-group processing in E-Miner

kdp — Tue, 28 Jul 2009 13:59:04 GMT

Thanks miner40!

I checked E-Miner and I am afraid I don't have those nodes. I am currently using E-Miner 5.2 which is probably why I don't see them.

Anyways, this will be of good help when we upgrade to the new E-Miner.

kdp