This section introduces SQL statements that can be used
for Intelligent Miner® Scoring.
Figure 1 maps the conceptual objects to the DB2® user-defined types (UDTs) used in Intelligent Miner. This figure shows an example for clustering.
In Figure 1 the
Scoring function DM_applyClusModel receives the
following parameters:
- A PMML model of type DM_ClusteringModel.
- A database row encapsulated in an object of type DM_ApplicationData .
The REC2XML function creates an XML string value representing the database row, because this parameter contains a list of column name/value pairs for all columns in a row.
As the scores are computed in a scalar user-defined function, the scoring step can be embedded in arbitrary SQL statements.