AI GLOSSARY

Data Drift

Data drift occurs when the distribution between the inputs of a machine learning model change, which can cause a model to perform poorly. This shift happens when the characteristics of the data used for training the model differ from the data it encounters in production. Monitoring and addressing data drift is crucial to ensure that models continue to make accurate predictions in real-world applications.