Setting up the Project

After completing the initial configuration, set up a project from the AI Matching UI in ibi MDM.

Before you begin

See Preparing for AI Matching.

    Procedure
  1. In ibi MDM, click Administration > AI Matching.
    Note: If the AI Matching configuration is not loaded on the UI, click Restart.
  2. On the Setup Selected Project tab, click Create and confirm the action.
    The new project is listed on the Projects tab.
    The project directory is created under the specified directory mentioned in the LearnConfig.xml file under the RootDirectory tag. The project directory signifies all active and deployed learn models.
  3. To export the active project to the Edit directory mentioned in the LearnConfig.xml file, click Edit and confirm the action if prompted.
    The project is available for editing. If the project is in the edit mode, you cannot perform any other action such as deploy and approve. For details, see AI Model Process.
  4. Open the editable project in (Undefined variable: MDM-global-variables.prod_patterns_learn_second) and define features.
    A feature is a query or querylet defined on the fields or group of fields used as matching criteria in the record matching process. For example, to search for records based on the first name, last name, or the street address, define feature first name, last name, or the street address as fields.
    Restriction:
    • For child repositories, you can define only one feature in (Undefined variable: MDM-global-variables.prod_patterns_learn_second).
    • Do not merge attributes or features from parent and child repositories.
    • Do not create attributes with the same name in parent and child repositories.

    For information about defining features, see ibi Patterns - Search Learn UI Guide.
  5. On the Projects tab, click Complete Edit and confirm the action if prompted.
    The project state is changed to Features Defined in the Create State column.
What to do next

After you set up the project, see Training an AI Model.