Research on Physical Education Teaching in Universities Based on Deep Learning

Abstract

In this paper, we utilize deep learning methods to perform automatic feature learning from multi-source heterogeneous data, mapping different data into the same hidden space, and obtaining the deep features of the data associated with students and sports. A small amount of information is filtered from a large amount of input information by assigning weights in the attention mechanism and higher weights are assigned to it, and a collaborative filtering recommendation algorithm based on deep learning and algorithm evaluation index is proposed. The deep learning algorithm is integrated into the classroom of “basketball one-handed in situ over-the-shoulder shooting” in college B for empirical analysis, and the teaching effect is compared with that of the traditional teaching mode. The results show that the average scores of 50 meters, sit-ups, pull-ups, 800 meters, 1,000 meters, seated forward bending, and cross-direction running of the students in the physical education classroom based on deep learning are higher than those of the students in the traditional physical education classroom by 5.94, 6.83, 7.69, 6.56, 5.87, 5.1, and 3.68 in the physical fitness comparison analysis. Study shows that physical education teaching based on deep learning can improve students’ various physical qualities, which is of great significance and value in accelerating the high-quality development of physical education and promoting students’ comprehensive development.

Keywords

• deep learning
• attention mechanism
• assigning weights
• collaborative filtering
• physical education
• 47n50