Research on real-time data analysis system of sports training based on reconfigurable data mining technology

Abstract

This paper utilizes smart wearable devices to quickly capture, record, and store performance information (calorie consumption assessment, pedometer, and heart rate monitoring), location information, and biochemical data of sports training populations through the sensory layer. Based on the reconfigurable data mining-related technology, combined with the data and test scores of the sports training population, data mining analysis is carried out to form valuable physiological, physical, and other characteristic information, which provides a scientific and referable strategy for sports training, thus also achieving the purpose of improving the training effect of athletes. The results show that, from the visualization and analysis results of the decision tree model for boys and girls, the girls’ long-distance running event (800 m) and boys’ pull-ups are located in the root node of the decision tree, which has the most significant effect on the performance level. Consequently, sports training programs were created that were more logically based on the distinction between male and female students. In individual exercise energy expenditure, by 420s, the limiting energy expenditure rate of 5.6 kcal/30s exceeded that of the athletes. Moreover, with time, energy expenditure continues to decline. The accuracy of this test can be 90%.