Scoring using a modeling stream

In an IBM® Campaign flowchart, open an SPSS Score process box to configure the score settings. These settings are used when the audience members are scored during an IBM® Campaign automated process flow.

You can modify the scoring options in an existing modeling stream. For example, you might want to take a snapshot of the values that were used to generate the score, such as the account balance at the time the score was derived. These fields might be used for personalization, such as sending the message "Thank you for your recent deposit of $500."

The configuration property Campaign | partitions | partition[n] | server | internal | numRowsReadToParseDelimitedFile is used by the Score process box when importing a score output file from IBM® SPSS® Modeler Advantage Marketing Edition. This property determines how many line entries to examine in order to determine data types and field lengths. For more information, see the IBM® Campaign Administrator's Guide.

To score using a modeling stream, complete the following steps.

In IBM® Campaign, open a flowchart in Edit mode.

Drag process boxes from the palette to the flowchart workspace. Connect them as you normally would to select a set of audience members.

For example, use the Select, Segment, and Merge process boxes.

Drag the SPSS Score process box from the palette to your flowchart workspace.

Connect the process box that contains the audience members to be scored to the SPSS Score process box.

If any of the modeling streams in the SPSS Score process are created in the same flowchart, connect those SPSS Model process boxes to the SPSS Score process box. Create temporal connections to ensure that scoring does not occur until the model is refreshed. You cannot use the temporal connection if the model has not already been built. To create the temporal connection, drag one of the arrow corners of the SPSS Model process box to the SPSS Score process box.

Double-click the SPSS Score process box in the flowchart workspace.

The SPSS Score Process Configuration dialog opens.

Select the source cell that contains the audience members to score.

Select the number of models to use for scoring.

Configure each model by completing the following steps:

Select the modeling stream either by browsing to a file in the repository or by uploading a modeling stream from your local hard disk drive. After you select a valid modeling stream, the scoring outputs are displayed in the Score Field Name column.

You cannot configure the SPSS Score process if the model is not already built.

Optional: In the Parameter For column, change the default model name to a more user-friendly name. The downstream scoring outputs are identified by this model name.

For example, you can name the output fields $XS-Response and $XSC-Response for the output score and associated confidence value. The model name is FreeShippingOffer. The scoring outputs available downstream from the SPSS Score process are named FreeShippingOffer.$XS-Response, for the free shipping model score, and FreeShippingOffer.$XSC-Response, for the associated confidence value.

Optional: Select a modeling stream row and click Match Variable Names to verify and match any incorrectly matched fields.

By default, variables are automatically matched by name. Verify that the matches are correct and manually fix any incorrect matches. Auto-matching is described in Automatic matching of fields.

Optional: Click Open Modeler Advantage Session to configure the scoring options.

IBM® SPSS® Modeler Advantage Marketing Edition opens in a new window. You can change the default scoring outputs (or any other scoring options) and save the modeling stream.

Click OK and close the SPSS Score Process Configuration dialog.

If IBM® Campaign detects that the SPSS Score process is not completely configured, the SPSS Score process box displays the status unconfigured with an error message.

10.

Optional: Automate the flowchart to run periodically. You can either use the Schedule process in your flowchart or configure the IBM® EMM Scheduler. The Scheduler is described in the IBM® Campaign User's Guide.

11.

Save the flowchart.

You can now perform a test run on the flowchart to verify that it returns the results you expect.