Configure class recall for a classification solution

  • Release version: Xanadu
  • Updated August 1, 2024
  • 1 minute to read

    • Create and apply a class recall parameter to an ML solution prior to training its data. For example, you set and apply this solution parameter to 90% accuracy for all records you train in the Email class.

    Before you begin

    Note:
    Configuring advanced settings on your ML solutions is optional. If you choose to configure any of these settings, make sure you're well informed regarding the technology you're enabling in the solution, and that you have a use case that benefits from what the technology offers.
    • Create and save a classification solution definition or use an existing one.
    • Role required: admin or ml_admin

    About this task

    The class recall solution parameter enables you to steer a solution's training to bias a specific class. For example, classifying an incoming email as a Phish or not can be an important use case in a security-related machine learning solution. In this situation, it's very important to identify every Phish, and it may be okay to report a non-Phish as a Phish occasionally. However, no real Phish should be classified as a non-Phish. In such situations, the recall metric must have a high value, which might lead to lower percentages for precision and coverage.

    Procedure

    1. Navigate to All > Predictive Intelligence > Classification > Solution Definitions.
    2. Open a saved classification solution definition form.
      In this example scenario, you use an Incident Categorization solution definition form that you haven't trained yet.
      This image shows an example classification solution definition on which you apply the class recall parameter.
    3. On the Advanced Solution Settings tab in the Related Links section of the form, click New.
      This image shows how to select the Solution Parameters option for creating the parameter.
    4. Create a parameter record.
      1. In the Solution Parameters field, click the search icon.
      2. In the ML Solution Parameters screen, select Add class recall value while training.
        How to create the parameter record by selecting a Search value, and then selecting the Class-Recall key Short Description.
    5. Click Submit.
      The Advanced Solution Setting record appears.
      This image shows the Advanced Solution Setting record for Class Recall.
    6. Configure the User inputs field.
      1. Enter the class name on which you want to track the Recall value.
        In this scenario, you enter Phish for the ClassName.
      2. Enter the Recall value.
        Enter 90 for the RecallValue.
      Here you're specifying Phish as the target class, and 95 is the recall percentage you're requesting the system to deliver during solution training.
      How to configure the User inputs field, where Spam is the target class, and 90% is the prediction accuracy you're asking the system to deliver during solution training.
    7. Click Submit.

      Result: Class recall is configured for your classification solution. Its solution parameter appears on the Advanced Solution Settings tab of your classification solution definition form.

      When you submit the record you created, the class recall solution parameter appears on your clustering solution definition form.

    What to do next

    Train your saved classification solution.