topic Re: EM Data Source gives variables role=residual in SAS Data Science

EM Data Source gives variables role=residual

kah1 — Wed, 19 Sep 2018 18:34:19 GMT

I have a dataset that when that creates a role of "residual" vs. "input" for several variables when added as a data source in E-miner. I thought it might have to do with the % missing so I choose the Advanced Options and changed the percent missing threshold and ran but it did not change the role to input. How does the "residual" role get determined? I can manual change them to input but I want to understand if there is an option controlling this.

Thanks

Re: EM Data Source gives variables role=residual

WendyCzika — Thu, 20 Sep 2018 15:00:43 GMT

Do the variables marked as Residual start with the prefix "R_" ?

Re: EM Data Source gives variables role=residual

kah1 — Thu, 20 Sep 2018 15:43:59 GMT

Almost, the variables start with "RD_". Why is that a specific variable name I shouldn't use? If so what else should not be used?

Re: EM Data Source gives variables role=residual

kah1 — Thu, 20 Sep 2018 17:38:09 GMT

I renamed all my variables and not it makes the role ="input". Thanks let me know if I should be aware of any other prefixes I shouldn't use

Re: EM Data Source gives variables role=residual

WendyCzika — Thu, 20 Sep 2018 20:31:52 GMT

This post gives a list of other variables with reserved names or prefixes in EM:

https://communities.sas.com/t5/SAS-Communities-Library/Scoring-Series-Part-2-SAS-Enterprise-Miner-Scoring-Output/ta-p/223842

You can also take a look at this post about global metadata to see how to programmatically change variable roles, levels, etc. instead of manually:

https://communities.sas.com/t5/SAS-Communities-Library/Tip-Defining-Global-Metadata-for-SAS-Enterprise-Miner-Projects/ta-p/223835

Re: EM Data Source gives variables role=residual

kah1 — Fri, 21 Sep 2018 12:14:58 GMT

thank you, makes sense to me now!