Which of the following best describes the concept of 'model drift'?
-
The model's performance improves over time
-
The model's predictions become less accurate due to changes in data distribution
-
The model's architecture changes during deployment
-
The model is retrained with new data
