Edit Feature Extraction Query

To edit a a Feature Extraction Query Node:

1. Double-click the Feature Extraction Query node or right-click the node and click Edit. The Edit Feature Extraction Query Node dialog box opens.

2. In the Feature Extraction Query Node dialog box, enter the details in the following tabs:

   • Feature Predictions tab:

     • In the Case ID field, select a case ID from the drop-down list. The case ID is optional. It is recommended that you specify a case ID to uniquely define each record. The case ID helps with model repeatability and is consistent with good data mining practices.

     • In the Number of Features to Extract field, specify an input. The default is 10.

     • Edit Feature Prediction Outputs. Oracle Data Miner automatically determines the output for the query. You can modify the output.

   • Partitions tab:

     • Add one or more Partition attributes. This is optional. Selecting a Partition attribute directs the predictive query to build a virtual model for each unique partition.
       To add partitions, click in the Partitions tab. Use the Add Partitioning Columns dialog box.

     • Specify partitioning expressions.

   • Input tab: You can perform the following tasks:

     • Add input

     • Modify input

     • Change the mining type

   • Additional Output tab: You can add output. This is optional. By default, all target columns, the Case ID column, and partitioning columns are automatically added to Additional Output. To make changes, click Additional Output.

3. Click OK.