How ThingPredictor Works
 |
ThingPredictor software media is no longer available for download as of 8.3.
Any new model that you publish from Analytics Builder is uploaded to the Analytics Server instead of ThingPredictor.
|
ThingPredictor can deliver predictive data analytics in real time. It handles synchronous data scoring based on a pre-trained data model and a dataset from which you want to predict an outcome.
|
|
A new model is not required for every data scoring request. Once the prediction model is generated, it can be reused as often as necessary.
|
The following diagram illustrates the ThingPredictor process flow:
User Actions
1. Deploy the ThingPredictor agent.
2. Build a model in Analytics Builder.
3. Configure a provider in Analytics Manager.
4. Publish the model from Analytics Builder to Analytics Manager.
5. Process the model in Analytics Manager, including determining inputs and outputs, configuring data mapping, and testing.
6. Associate the model with a thing in ThingWorx Composer.
7. Subscribe a property of the thing to an event in ThingWorx Composer.
Automated Actions
1. Analytics Manager listens for the subscribed event.
2. Analytics Manager is notified when the event occurs.
3. Analytics Manager processes the event data and submits a request for scoring to the ThingPredictor agent.
4. ThingPredictor runs scoring on the submitted data and outputs results in a JSON format.
5. The results are returned to both Analytics Manager and the thing in Composer.