AI GLOSSARY

Concept Drift

Concept drift occurs when the relationship between the input data and the target value changes in some way, potentially making the model inaccurate or unreliable. Concept drift can lead to a decline in the model’s accuracy because it was trained on data that no longer reflects the current patterns or relationships. Handling concept drift is essential in dynamic environments, such as financial markets or user behaviour prediction, where conditions evolve continuously.