Updating Machine Learning references

You can update the deployment IDs or use a different MLz Deployment ID file to alter the input/output parameters of your Machine Learning enhancement.

About this task

You manage Machine Learning references from the Rule Designer project context menu. The Machine Learning references panel has three options:
  • Create: Import Machine Learning models; you can import as many as you like
  • Update: Update or refresh existing Machine Learning models
  • Delete: Removes Machine Learning models from the list

The following task focuses on the Update option of integrating your machine learning models.

Procedure

  1. Right click your rule project and select ODM for z/OS > Enhance with Machine Learning....
    The Update a Machine Learning call panel opens.
  2. Select an existing Machine Learning reference from the list.
  3. Select Update to open the Update your Machine Learning call panel.
    You can modify two fields in this panel: Machine Learning data and MLz deployment ID. If you update the deployment ID, all of the rules that reference the previous ID are updated to reference the new deployment ID. If you re-import from a YAML model, then the Java™ XOM and the BOM are updated to reflect the new (or removed) input or output variables.
    When you update a machine learning model, the Java XOM is completely re-created, with the exception of the //user content sections. The default class is also updated to do the following:
    • Insert new default output variables getters (returning null)
    • Remove default output variables getters, provided you have not modified them (that is, they still just return null).
    • Change the types on the getter classes. It does this even if you have customized getters to provide custom defaults.
    Errors are created in the files, because you can use the errors as reminders that you must update the returned values to the new types.