Embedded ML - Planned Shipment ETA Prediction

The new Embedded ML framework provides machine learning capabilities embedded within OTM – users can take advantage of the ML capabilities from within their production / dev environment.

The existing External Service (IoT IA) based shipment ETA prediction is now also available as part of the Embedded ML framework. With this feature, users now do not need to subscribe to IoT IA and OCI Object Storage to use shipment ETA prediction. All other functionality will remain the same.

Here's a look at what has changed. Notice that the Planned ETA Prediction Objective Model Type is now also available under the Embedded Learning Objective.

Before: uses the external ML service (IoT IA)

External Service (IoT IA)

Now: uses the embedded ML framework in OTM

Embedded Learning

In 24C, Embedded ML shipment ETA prediction only supports the basic end-end flow of the Planned model.

Users can create Projects and Scenarios, Load Data to Analytics, Perform Training, View Training Results, Perform Prediction, View Prediction Results and analyze their ML input data and Training/Prediction results on dashboards.

Filters, Outliers, Included/Excluded Columns and Process Management are not yet supported. In-transit ETA prediction in Embedded ML is not yet supported. They will be available in a future release.

Using the shipment ETA prediction feature is now easier since subscription to IoT IA and OCI Object Storage is not required.

Steps to Enable

You don't need to do anything to enable this feature.

Tips And Considerations

• This feature is only available to customers who are on ATP Database pods. If you are not on ATP yet, you will be shown a corresponding message when you try to use this feature.

• The existing external service (IoT IA) based ETA prediction solution will be phased out in a future release.