Hybrid neural network-based exploration on the influence of continuous sensor data for the balancing ability of aerobics students.

In the era of Machine Learning and Internet of Things, numerous sensing devices collect sensory information for various industries. Aerobics, a popular and energetic sport, benefits from analyzing and researching the training approaches of high-level university and college teams. However, the complexity and diversity of sensor data require systematic classification and adaptive preprocessing for effective human motion recognition. Based on these, this paper aimed at the balance ability of aerobics students and analyzed and studied the balance ability of aerobics students through wireless sensor networks. Furthermore, this paper provides a comprehensive overview of how to use a class of enhanced machine learning approaches, such as Artificial Neural Networks, to facilitate analytics and learning in the wireless sensor networks domain by utilizing continuous sensor data. During experimental work, this study selects 10 third-year students in a college of physical education, and the subjects were divided into two groups: the experimental group and the control group collecting their data by continuous sensors. The experimental group mainly carried out special training and elastic band exercises within eight weeks. After eight weeks of elastic training, eight indexes of the left and right legs, including anterior (ANT), anterolateral (ALAT), lateral (LAT), posterolateral (PLAT), posterior (POST), posteromedial (PMED), medial (MED), and anteromedial (AMED) were improved (P < 0.01), but the indexes with the left foot as the fulcrum were not significantly improved. By using the 8-week elastic band training method, the technical indexes such as single leg lifting and single leg rotation of aerobics students were significantly improved (P < 0.05). [ABSTRACT FROM AUTHOR]