Abstract: Concept drift, referring to the temporal evolution of data distributions, presents a significant challenge to the deployment of machine learning models in dynamic, real-world environments.