Teenager Physical Fitness Evaluation Model Based on Wearable Running Recordings

Abstract

In recent years, the fitness problem of teenagers have attracted people’s attention. The rapid development of biosensors and Internet of things technologies makes it feasible to use wearable devices for teenager mobile health monitoring. In this research, we collect 513 middle school students’ (288 boys and 255 girls) photoplethysmography signals while running with the aid of our customized smart handbands. We extract the time duration while running, increase rate of HR, HR rate reserve, descent rate of HR, maximum HR and SpO2 from the original recordings and propose a physical fitness evaluation model based on support vector machine method. The precision accuracy of our model reaches 99.02%. Our findings provide evidences supporting the possibility of evaluating teenagers’ health levels by wearable physiological recordings.