topic SAS-EM Data Source Columns Roles in SAS Data Science

SAS-EM Data Source Columns Roles

beckeybianc — Thu, 14 Dec 2017 08:53:12 GMT

Hi everybody,

I like SAS EM but i stucked.

Can you explain me datasource variable roles? When should i use them? And what is Drop, Report and Level?

Assessmet

Censor

Classification

Cost

Cross ID

Decision

Frequency

Input

Key

Label

Prediction

Referrer

Residual

Segment

Sequence

Target

Text

Text Location

Time ID

Treatment

Web Address

best regards

Re: SAS-EM Data Source Columns Roles

WendyCzika — Thu, 14 Dec 2017 19:02:41 GMT

In the SAS EM Reference Help (from Help>Contents in EM), go to Data Sources>SAS Enterprise Miner Sources, then the section "Model Roles and Measurement Levels of Variables" and it has the information on all of these variable roles. Drop, Report, and Level are explained in there too, but I'll include that here since that part isn't too big 🙂

• Level — is the measurement level of the variable. (MORE BELOW)
• Report — indicates whether a variable should be automatically included in reports such as the predictive model assessment decile and percentile tables.
• Drop — drops the variable.

Select one of the following measurement levels:
• Binary — contains two discrete values (for example, PURCHASE: Yes, No).
• Interval — contains values that vary across a continuous range (for example, TEMP: 0, 32, 34, 36, 50, 56, 80, ...., 102).
• Nominal — contains a discrete set of values that do not have a logical ordering (for example, PARTY: Democrat, Republican, other).
• Ordinal — contains a discrete set of values that do have a logical ordering (for example, GRADE: A, B, C, D, F).
• Unary — contains one discrete value.