What does 'model drift' refer to in the context of deployed models?
-
The process of updating the model with new data
-
The degradation of model performance over time due to changes in data distribution
-
The initial training phase of the model
-
The difference between training and testing datasets
