Enhancing college students physical education using artificial intelligence-optimized teaching system based on biomechanics

Article ID: 503

Abstract

Physical Education Teaching concerns the process of leading students to perform different tasks, games, and workouts that improve physical fitness, body control, and health. In the realm of cell and molecular biomechanics, physical education teaching can be regarded as a means to induce specific physiological responses at the microscopic level. The various physical activities in which students partake, like diverse tasks and workouts, exert mechanical forces that permeate throughout the body and impinge upon cells and tissues. During physical exertion, cells within muscles, bones, and connective tissues are subject to biomechanical stress. This stress triggers a cascade of molecular events. Teachers focus on enhancing spatial and manual skills, promoting cooperation, and setting up priorities. In this research, it is proposed to learn about the teaching system of physical education in colleges and universities using artificial intelligence (AI) optimization algorithm. Thus, for predicting the achievements of college students in physical education, we propose the Blue Monkey optimization-driven Weight-Tuned AdaBoost (BM-WTAdaBoost) algorithm. The observations and variables were derived from typical physical education programs of college students during their training sessions. A data pre-processing technique known as min–max normalization is applied to the obtained raw data to enhance its quality. For nonlinear data, Kernel Principal Component Analysis (kernel-PCA) is employed as it helps in extracting the nonlinear information, which in turn helps in making accurate predictions. The following is our proposed model: BM opt with WTAdaBoost to improve selecting features and model accuracy in predicting college students’ physical education outcomes. Python program uses our suggested technique. The finding assessment phase assesses the suggested model’s prediction efficacy using several measures, including the accuracy ratio (99.8%), F1-score (95.56%), prediction ratio (98.24%), interaction ratio (97.2%), efficiency ratio (98.24%), performance ratio (97.2%), and error rate (5.62%). We also performed a comparative analysis with different traditional approaches to assess the efficacy of the suggested strategy. Comparative analysis with traditional methods shows the superiority of this approach in predicting physical education outcomes considering cell and molecular biomechanics, providing a novel perspective for understanding and optimizing physical education in relation to the microscopic biological world.