Skip to main content

Table of Contents

Configuring Modeled Analysis


Only available versions of this content are shown in the dropdown

You can configure your Analysis Project to use modeled analysis. This approach uses statistical analysis to identify properties that have a high influence on the targeted outcome, completely removing the need for guess work or intuition when assembling criteria in a Segment.

A key aspect of this analysis approach is the use of a closed loop feedback mechanism. Sample data is used to train or drive the creation of modeled analysis predictors. The sample data brings with it details of previous interactions and responses to actions made. While we are creating modeled analysis predictors that have predictive power over various outcomes, they are based on positive and negative responses to previous actions.

  1. Click the Modeled Analysis tab of an Analysis Project.

  2. In the Select action section, you can train the analysis on specific instances of responses to previously delivered actions. You can train the analysis using responses to previous actions for the same product or product group. Alternatively, for a new product offering, you can use similar products or product groups to assemble likely predictors that will help illustrate the type of customers who are likely to buy the new product.

    1. Select the Issue, Group, or Action on which you want to focus.

    2. Leave any of these values blank to bring through all entries for the parent category. For example, if you select the Sales Issue and the Handsets Group, and leave the action field blank, the analysis will consider the responses to all actions that belong under Sales or Handsets.

  3. In the Interaction Details section, you can train the analysis on specific interaction details, in a similar way as with responses to actions. Select the Direction, Treatment, and Channel on which you want to focus. Use the Offered After field to restrict the analysis only to responses to those actions that were offered after a specified date and time.

  4. In the Select Predictors section, remove or add properties from the base Sample which will be used as predictors for the analysis. Predictors which have no predictive value in modeled analysis should be removed from this section.

  5. In the Define Behaviors section, select positive and negative outcomes for the modeled analysis based on the set of responses received so far.

    Review each response type in the Available Behaviors section and mark it as a positive, negative, or ignored behavior for the analysis. To reload the list of available behaviors, click Refresh. Marking as Positive or Negative moves the response type into the corresponding response category (Positive Behavior or Negative Behavior).
Did you find this content helpful?

Have a question? Get answers now.

Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.

Ready to crush complexity?

Experience the benefits of Pega Community when you log in.

We'd prefer it if you saw us at our best.

Pega Community has detected you are using a browser which may prevent you from experiencing the site as intended. To improve your experience, please update your browser.

Close Deprecation Notice
Contact us