Once I published the model to ThingPredictor, and use Thingworx Analytix Managemer to test new data with ThingPredictor connector. Does the predictor use its own engine or use Thingworx Analytics Server to calculate the result?
Hi Zhuwenxi
ThingPredictor includes its own scoring service so it does not require ThingWorx analytics Server at all for runtime scoring.
The only time you may need ThingWorx Analytics Server with ThingPredictor is to define the model:
- If you are using an external model imported through pmml file, there is no need of ThingWorx analytics Server.
- however if the model is published from ThingWorx Analytics Server, then obviously at this time you need it but only for the Publish operation to create the model.
Once the model is created, there is no more need for ThingWorx analytics Server. ThingPredictor will score on its own.
See https://www.ptc.com/en/support/article?n=CS265570 for some info about the different Analytics products.
Note that this is true up to release 8.2 (the current one).
Hope this helps
Kind regards
Christophe
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
