Concept drift refers to changes in?

• A. Changes in input data distribution over time
• B. Changes in the relationship between inputs and outputs over time
• C. Changes in model hyperparameters
• D. Data storage changes

Answer: (B)

Concept drift happens when the relationship between the features you use and the target you’re trying to predict changes over time. In other words, the mapping from inputs to outputs evolves, so a model that once performed well may become less accurate unless it is updated or retrained to reflect the new relationship. This is different from shifts in the data you observe (input distributions changing), which is data drift rather than a change in the underlying concept the model is trying to learn. Changes in hyperparameters or in data storage don’t alter the predictive relationship learned by the model, so they aren’t the drift concept being described.