Predictive Intelligence Workbench pretrained use cases

The pretrain functionality mechanism for use cases enables and trains the machine learning solution definition when you activate the Predictive Intelligence Workbench application.

When a pretrained machine learning solution exists in the system, Predictive Intelligence Workbench highlights the use case templates with the Pretrained flag.

Figure 1. ITSM Predictive Intelligence Workbench pretrained model use case template

The Pretrained flag for use case templates displays when the following criteria are met.

• The Predictive Intelligence Workbench piwb.instance_eligible_auto_train system property for pretraining is enabled (true) on the your instance. This property is disabled by default. Requires the piwb_admin role to enable.
• The instance has to be one of the following types:
  • ded-prod
  • shared-prod
  • ded-subprod
  • shared-subprod
• The associated machine learning solution is active and ready for deployment.
• The use case template does not contain an existing pretrained use case.
• The machine learning solution net automation is greater than the threshold value.
  Note:

  The default base-system threshold value is 30. You can configure this value via the Predictive Intelligence Workbench piwb.auto_train_threshold system property for. Requires the piwb_admin role to modify the default value.
• The machine learning solution is a pretrained base-system solution with an associated record for its definition in the ML Auto Train Solution [ml_autotrain_solution] table.