Configure the Machine Learning Service Classification Prediction
The Machine Learning mApp® Solution contains a machine learning model, supplemental Incident form, expressions, scheduled item, and associated One-Step™ Actions.
To configure the Machine Learning mApp Solution:
- Determine who in your organization will act as machine learning subject matter experts (SMEs) for reviewing the service classification predictions. This could be higher-level service technicians or managers.
- Place the Verify Incident Confidence, Reclassify Incident for Confidence, and Re-Predict Classification links on your Incident Overview form. These fields are available in the Machine Learning Supplemental form included with the mApp Solution file.
- Set view and execute permissions for the Verify Incident Confidence and Reclassify Incident for Confidence links on the Incident form.
- Set view permissions for the Machine Learning dashboard.
- Publish the .mApp Blueprint.
- Set up the Scheduled Item. For detailed instructions on Scheduled Items, see Create a Scheduled Item.
- Provide general and schedule information for the Scheduled Item.
- On the Action page, select Train Machine Learning from the Action menu.
- Select the ellipsis button by the Machine Learning Training Configuration field to open the Machine Learning Manager.
- In the Association field, select Incident.
- Double-click the Incident Subcategory machine learning model.
- For the Training Query, select Total Incidents and Requests with High Confidence (Global > Machine Learning).
Before you run the first training scenario, review outstanding tickets to classify them as high confidence or add a condition to the search to include any items you want in the training group.
The configuration details autopopulate.
- Allow the included Scheduled Item to run.
Let the Scheduled Item run as scheduled; running it by using the Test button in the Scheduled Item Manager won't work.
Create a Scheduled Item to run the machine learning model.