Integrating a SAS Enterprise Miner Model with
Existing SAS Data
Overview
You can use a Mining Results transformation to integrate a SAS Enterprise Miner model
with data sources in your SAS Data Integration Studio data warehouse. Using the
Mining Results transformation enables you to associate a SAS Enterprise Miner model
with a job and use that job to create an output table that applies the model to the source
data.
Note: One or more SAS Enterprise Miner models must be registered in your metadata
repository before you can use the Mining Results transformation.
Problem
You want to create a job that creates a target table from a SAS Enterprise Miner model.
700 Appendix 3 Miscellaneous Transformations
Solution
You can use the Mining Results transformation to create a job that creates a target table
from a SAS Enterprise Miner model.
In this example, a statistician uses SAS Enterprise Miner and historical home equity data
to build a data mining model to predict if a customer might default on a home equity
loan. After the statistician builds the model, it is registered in a SAS metadata repository
that a SAS Data Integration Studio developer can use. Additional customer data is
collected using SAS Data Integration Studio. The new data has the same customer
information but does not contain the predictions about a customer's probability for
defaulting on a home equity loan. The SAS Data Integration Studio developer applies
the SAS Enterprise Miner model to the new data source to generate the prediction of the
customer's probability to default on a home equity loan.
To use the Mining Results transformation, your SAS Enterprise Miner models must be
registered in the same metadata repository that contains the sources for your job. You
can use the SAS Enterprise Miner Configuration Wizard to associate your SAS
Enterprise Miner metadata repository with the metadata server that is used for your SAS
Data Integration Studio application. See the SAS Enterprise Miner online Help for more
information about how to use this wizard.
Note: It is recommended that you group your models in trees to make them easier to
find when using the Mining Results transformation.
For best performance, the target of a Mining Results transformation should have only
those columns that are required. These columns include the required input variables form
the source table and the output results that are specified in the SAS Enterprise Miner
model.
Perform the following steps:
“Create and Populate the Job” on page 701
“Associate the SAS Enterprise Miner Model with the Job” on page 702
“Run the Job and View the Output” on page 704
Tasks
Create and Populate the Job
Perform the following steps to create and populate the job:
1. Create an empty SAS Data Integration Studio job.
2. Select and drag a Mining Results transformation from the Transformations tree in the
Data Transforms folder. Then, drop it in the empty job on the Diagram tab in the
Job Editor window.
3. Drag the cursor from the source table to the input port of the Mining Results
transformation. This action connects the source to the transformation. The following
display shows the process flow for the sample job.
Display A3.25 Mining Results Process Flow Diagram
Integrating a SAS Enterprise Miner Model with Existing SAS Data 701
Note that the source file for the sample job is named HMEQ.
Associate the SAS Enterprise Miner Model with the Job
Perform the following steps to associate the SAS Enterprise Miner model with the
Mining Results transformation:
1. Display the transformation properties window.
2. Click the Mining Results tab.
3. Select the data mining tree model. Note that the Mining Results folder always
contains all tree models, regardless of their assigned trees. Therefore, you can always
find all available trees in the folder. Metadata about the selected tree model is
displayed on the tab. The tree model used in the sample job in HMEQTree. The
Mining Results tab for the sample job is shown in the following display.
Display A3.26 Sample Mining Results Tab
4. Click the Model Attributes tab. Use this tab to view the required input variables and
output results that were defined for the model in SAS Enterprise Miner. You can also
click View Source Code to view the SAS source code that is generated by the
model. The Model Attributes tab for the sample job is shown in the following
display.
702 Appendix 3 Miscellaneous Transformations

Get SAS Data Integration Studio 4.9 now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.