Data Drift

Overview

In our Drift overview, we describe the different types of drift commonly referenced and measured for a robust machine learning monitoring capability. Data drift refers to a measure of similarity between production data and the model's training dataset.

🚧

Note

Data drift is only available for models that have data drift configured within the model container

Data drift object

JSON

{
  "modelIdentifier": "...",
  "modelVersion": "..."
}