Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

384 Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

10.3.1 About SAS Enterprise Miner

SAS Enterprise Miner is an extensive toolset that supports and enables all of the

stages of the data mining process that we have discussed in this chapter.

Figure 10-14 on page 385 shows an example of the SAS Enterprise Miner

workbench.

The modeling GUI looks similar to the Design Studio in InfoSphere

Warehouse 10.1 and is composed of data mining flows that process and pass

data from one operator to another.

SAS Enterprise Miner provides a robust data mining environment for

򐂰 Data exploration

򐂰 Data preparation

򐂰 Prediction

򐂰 Model comparison

In a traditional SAS Enterprise Miner workflow, processing and usage proceed

as described here:

1. Extract a test sample data set from some source (for example, InfoSphere

Warehouse). Use that test set to develop the predictive data model with SAS

Enterprise Modeler.

2. Extract production data from data warehouse sources such as InfoSphere

Warehouse.

5

Derived from a presentation “Change the game with in-database analytics: how to run SAS scoring

models inside DB2” by Eileen Lin, Mattias Nicola, and Subho Chatterjee, that was given at the 2011

IBM Information on Demand® conference.

Chapter 10. Techniques for data mining in an operational warehouse 385

3. Run the scoring model developed in SAS Enterprise Modeler on that

extracted production data set.

Figure 10-14 Example of the SAS Enterprise Miner workbench

This process is repeated as often as necessary, depending on data latency

requirements and the rate of input of new data into the warehouse. This process

is therefore problematic for the following reasons:

򐂰 Scoring is done on an extract of the data, which might already be out of date

at the time of the execution.

򐂰 Scoring is processed in an outboard processor instead of near the data in the

database engine.