Handling missing values in data streams: An overview

Missing values are a common problem in streaming scenarios, mainly due to equipment faults, network errors, and data unpredictability. This paper presents an overview of handling missing values in data streams, elucidating key concepts and summarizing recent studies that tackle this issue. It highlights limitations related to data stream requisites, concept drift exploration, and missing mechanism assumptions. Our discussion aims to indicate open issues and contribute to new research initiatives in this area.