Predictive Model Endpoints

A model endpoint is a connection between the models you have trained in a workspace and where you make predictions. Without an endpoint, you cannot make predictions. If you have multiple models in the same workspace, you may want to run predictions on both simultaneously, or maybe you just switch back and forth regularly, in this case, you will want to create a new endpoint.

If you have a new model built in your workspace and you want to switch your current model endpoint to your new model you can find more information here.

Deploy a New Application

A new model endpoint can be created inside of the workspace you are using by going to the Applications tab, and clicking "Deploy a New App".

New Application Window

From here you can create several new types of applications, but for now, we are interested in creating a new predictive model endpoint. Anomaly detection models also use predictive model endpoints despite the name.

New Model App

Then we can give the new model a name and description, clicking "Continue" will bring you to the next step which will be choosing the model the predictions will go through, and selecting where the predictions will be saved.

Configuring the Model Endpoint

In the next stage, the first step is to select the model which can easily be done by clicking the Predictive Model line and selecting from the drop-down menu. If you choose to you can select the warehouse and table where the predictions will be stored, but this is not necessary. More information on data warehouses can be found here. Once you have done this click "Finish" and you can begin making predictions with your new endpoint.

Configure Model Endpoint

More information on making predictions here.