A Smart Architecture Leveraging Fog Computing Fusion and Ensemble Learning for Prediction of Gestational Diabetes

Abstract

Gestational diabetes (GD) is a growing global concern, underscoring the need for early detection and effective management to prevent adverse health consequences. This paper presents an innovative and reliable architecture to predict gestational diabetes in pregnant women. While reducing the frequency of doctor visits by sending the necessary data via Internet of Things (IoT) technology and receiving the prediction results via a mobile application in real time. The proposed architecture is a fusion of fog computing hardware with ensemble machine learning to enable low-latency, energy-efficient solutions for data processing, and cloud computing. The GD_Fog architecture leverages fused fog computing and load balancing techniques to reduce latency, power consumption, Network bandwidth consumption, and response time, and cloud computing is used based on the concept of use on demand for more reliability while harnessing the power of group learning to improve prediction accuracy. In addition, GD_Fog can be configured for different operating modes to ensure optimal quality of service and prediction accuracy in various fog calculation scenarios, which meet different user requirements. Through extensive testing using real-world data from pregnant women, the framework shows promising results, outperforming the latest methods in accuracy and efficiency. Where the percentage of improvement in prediction accuracy was approximately 6.5% when using ensemble learning, and the improvement in energy use, amounted to approximately 87.01% when using fused fog computing instead of cloud computing. These results confirm the potential of the proposed structure as an invaluable tool for the early detection and effective management of gestational diabetes.