Creating a predictive scoring agent project

Create a Predictive Scoring Agent project that contains a partly completed code template that you can populate.

About this task

You use the Predictive Scoring Agent wizard to create a predictive scoring agent project. The wizard also generates a code template that you then complete to finish your agent implementation.

Procedure

  1. Click File > New > Predictive Scoring Agent Project.
  2. Enter a name for the new agent, and select the solution project to which you want to attach the new agent.
  3. Specify an agent name and a bundle symbolic name, and then click Next. The bundle symbolic name must comply with OSGi naming conventions. By default, the wizard suggests to prefix this name with the name of the solution project to ensure its uniqueness and avoid conflicts between solutions with similar agents. Typically, the symbolic name follows the package names style syntax, for example com.example.agent.
  4. On the Connect to SPSS® Collaboration and Deployment Services page, specify the URL, user name, password if necessary, and click Next. If the SPSS Collaboration and Deployment Services server does not authenticate users, a user name and password are not necessary.
  5. On the Select Scoring Configuration page, select the scoring configuration to be called by the Predictive Scoring agent.
  6. Click Next.
  7. Enter a name for the scoring endpoint.
  8. Click Finish.

Results

The predictive scoring agent project is created.

What to do next

  • Complete the code template in the Java™ source file that has been created within the agent project, by mapping the inputs and outputs of the scoring model. You might want to obtain the values to map to the inputs from the event, from your entities or from aggregates, using the Java APIs. You might also want to store output values in a field within the bound entity.
  • Complete the agent descriptor .adsc file.
  • Add properties in the solution properties file. The solution_properties.xml file is in the root folder of the solution project. For more information, see Adding solution properties for predictive scoring agents. The property names depend on the value of the endpoint attribute of the @ScoringConfiguration annotation. Add the following scoring properties:
    • The HTTP or HTTPS URL of the scoring service.
    • The user ID authorized to access a secured scoring service. Specify the empty value if the scoring service is not secured.
    • The password for accessing a secured scoring service. Specify the empty value if the scoring service is not secured.
Parent topic: Creating predictive scoring agents