A real-time heartbeat monitoring using wearable device and machine learning

Wearable devices and edge computing enable for self-monitoring of heartbeat conditions. The advantage of self-monitoring is allowing for independent, anywhere, and anytime inspections. Typically, a wearable device on the market comes with a smartphone-based application and is meant for fitness tracking. Moreover, the device's ability to produce a gold standard cardiovascular recording enables the utilization of a sensor to capture cardiovascular data. This research provides a system for real-time monitoring and interpreting RR Interval data from Polar H10 data by using numerous machine learning methods. Therefore, the analyzer was trained to classify data into five categories: normal, supraventricular, ventricular ectopic, fusion, and unknown. The analyzers can predict heartbeats in less than one second, with the decision tree algorithm being the fastest to predict and the support vector machine algorithm is the most accurate.