Research on Wearable Sensors Data Collection and Health Assessment based on Machine Learning

With the rapid development of economy and culture, more and more people are starting to use wearable devices for health monitoring. However, as the health monitoring industry is still in its infancy, there are many problems in the professional knowledge, skill reserves, equipment use and logistics support of users. These issues lead to suboptimal health monitoring and can pose a threat to the health and safety of users. In this work, the health assessment of wearable sensor data acquisition and machine learning studied the discrimination method of integrating multi-source sensor signals and expert experience, and designed a two-stage anomaly recognition scheme. In the first stage, the discrimination rules containing expert experience were used to complete the preliminary discrimination of anomalies, and in the research of expert knowledge base, the representation and use of expert experience in wearable systems were discussed, and the construction method of anomaly discrimination knowledge base based on problems was designed and implemented. The Transformer model is used to process multi-source sensor signals, and its self-attention mechanism and timing expression characteristics are fully utilized, which greatly reduces the false negative rate of abnormal states. Experimental analysis shows that the proposed method significantly improves the accuracy and reliability of health assessment.