IoT and Machine Learning-Based Hypoglycemia Detection System

Hypoglycemia is a disorder that can be fatal at times. So, hypoglycemic event detection is necessary for avoiding critical situations and saving lives. Receiving the benefits of IoT and machine learning, a system has been introduced for hypoglycemia detection. The system is based on IoT that transfers data over a network without any human interactions. IoT sensors, Fog and cloud technology, and machine learning are integrated with this system. The system alerts the users before the condition becomes critical. Sensors maintain communication to Fog and cloud with TCP/IP protocol. The sensor’s data is transferred to Fog and then to the cloud for further analysis. A machine learning algorithm retrieves the data, analyzes, detects hypoglycemia, and generates necessary decisions. State-of-the-art machine learning algorithms such as XGBoost, Random Forest, KNN, and SVM were utilized for several analyses and comparisons. From the analysis, 94% accuracy was gained using the XGBoost that beat other algorithms.