Workflow for the Predictor Resource

The following figure illustrates the workflow for building, viewing, analyzing, and deploying analysis models.

1. Update user profile definition using the Profiles resource.
2. Import data into the Commerce Server Data Warehouse using the Commerce Server Data Transformation Services (DTS) tasks.
3. Add model configurations for your site to the Predictor resource using the Predictor Model Configuration Setup Wizard.
4. Build analysis models using the Build Model dialog box in the Predictor resource.
5. View and analyze a Prediction model in the Prediction Model Viewer.
6. View and analyze a Segment model using the Segment Viewer in Commerce Server Business Desk.
7. Deploy the analysis model onto your Web servers using PredictorClient API calls. For instructions, see Deploying Predictor.
8. Export user lists from segment and prediction models to the Business Desk List Manager module:
   • Use the Business Desk Affinity Lists module to export lists from prediction models.
   • Use the Business Desk Segment Viewer module to export lists from segment models.
   • Use the user lists to send direct mail campaigns.
9. Schedule and automate model rebuilding using the Model Builder Data Transformation Services (DTS) task in the Data Warehouse.

Note

• Prediction models provide more accurate recommendations than Segment models. Use Prediction models to provide recommendations on your site.

See Also

Installing Commerce Server Resources as Clustered Resources

Predictor Resource Tutorial

Securing a Predictor Deployment

Deploying Predictor

Setting Up Model Configurations

Predictor Configuration Model Setup Wizard

Copyright © 2005 Microsoft Corporation.
All rights reserved.