Key points are not available for this paper at this time.
Abstract Purpose The surge in Internet-of-Medical-Things (IoMT) and smart devices has resulted in a substantial influx of data streams within the healthcare domain. The interlinked structure of medical devices presents a pivotal hurdle referred to as Concept Drift, particularly significant in the medical arena due to the inherent instability of data patterns over time. In medical contexts, this complexity is heightened as sensors need to seamlessly shift from routine healthcare monitoring to managing urgent scenarios such as emergency ICU operations. The intricacy deepens owing to the uneven distribution of data in e-health scenarios. The complexity is further compounded by imbalanced data distributions in e-health scenarios. Method To address this challenge, our study proposes a novel Adaptive Ensemble Framework (AEF-CDA) specifically designed for detecting and adapting to concept drift in large-scale medical data streams from IoMT. The framework integrates adaptive data preprocessing, a novel drift-oriented adaptive feature selection approach, learning of base models, and model selection adapted to concept drift. Additionally, an online ensemble model is incorporated to enhance concept drift adaptation. Results The proposed AEF-CDA framework is evaluated using three public IoMT and IoT datasets. The experimental results demonstrate its superiority over contemporary methods, achieving a remarkable accuracy of 99.64% with a precision of 99.39%. These metrics surpass the performance of other approaches in the simulation. Conclusion In conclusion, the research presents an effective solution in the form of the adaptive ensemble framework (AEF-CDA) to effectively address the challenges posed by concept drift in IoMT data streams. The demonstrated high accuracy and precision underscore the framework's efficacy, highlighting its potential significance in the dynamic landscape of medical data analysis.
Maheshwari et al. (Mon,) studied this question.