Populate a data lake of X2 KPIs equivalent to that which will be received via E2.
Divide data into two sets: one set used for training the model, the other used for testing the trained model
Train QoE Predictor ML Model in a lab environment that has access to these metrics.