Siebel Smart Answer Administration Guide > Configuration and Implementation of Siebel Smart Answer >

Creating and Training Your KB

After you successfully install Banter Workbench, and create and categorize your corpus file, you are ready to create and train your KB.

Perform the following tasks in the sequence shown:

1. Create and analyze your KB in Banter Workbench.

   If you want to create a hierarchical KB, use the KB Editor to define the structure and use this KB to apply training and analysis on.

   When you use Banter Workbench to create and analyze in the same process, the main corpus is divided into two parts, as follows:

   • Training corpus, which is used to create the KB.
   • Testing corpus, which is used to analyze the performance of the KB when the KB classifies data.
2. Train and test the KB within Banter Workbench.

   In general, if you increase the number of items that you use for training the KB, the performance of the KB improves.

   As a final check before you introduce your KB into a production environment, use the Create, Analyze and Learn wizard, and the option called Create using all selected, analyze using all selected. Inspect the results and check the score for each item to verify that the results are accurate.

   After you begin using a KB file to categorize incoming email and requests, and agents select categories or give feedback, the feedback is stored in the KB file on the Siebel File System.

   CAUTION:  After you begin to receive feedback on a KB file, do not train that same KB file again. If you do, you delete all the statistical learning information based on your feedback.

3. Review Banter Workbench graphs and reports to analyze training results.

   The Cumulative Success Graph provides an overall view of KB performance. The Precision vs. Recall graphs help to determine category thresholds for generating automatic responses. For more information about thresholds, see Modifying a Category.