Engagement Prediction

Algorithm
Use Cases

Prerequisites

Before you start using our Recommender algorithm, ensure you’ve:
  1. Created a schema
  1. Integrated code to process your data and installed Begin’s Android SDK
  1. Created a project with the objective “Predict User Engagement”
Learning usually takes about 30 minutes. Once complete, your machine learning algorithm is ready to be used on your mobile device, browser, connected vehicles, or server applications.

Procedure

In order to infer a user engagement with an object, follow the initialization step:
import ai.begin.android.core.BeginWorker; BeginWorker beginWorker = new BeginWorker("APP_ID", "LICENSE_KEY", getApplicationContext());
To use your trained algorithm, input the following code:
BeginWorker.getInstance().predictEngagement("project_id", "user_id", "target_object_id", new ResponseInterface() { @Override public void success(JsonElement obj) { Log.d("Begin API Response", obj.toString()); } @Override public void error(Object obj) { Log.d("Begin API Response", obj.toString()); } });
This will return a predicted interaction between the provided user and the object. The possible interactions are:
  • INTERACTS_POSITIVE
  • INTERACTS_NEGATIVE
  • INTERACTS_NEUTRAL
  • NOT_FOUND (in the case one of the parameters provided is not available)
 
If you performed the initial processing in Python, you can still use the algorithm with other SDKs like Android or iOS SDKs.