Evaluation of physical education classes in colleges and universities using machine learning

Abstract

The traditional classes in colleges and universities are no longer fitted to the distinctive requirements of our society in terms of teaching, research, curricular settings, teaching resource allocation, and teaching quality evaluation. Machine learning approaches based on Internet technology are increasingly being introduced into classrooms and implemented across colleges and universities around the globe. To address the limitations of existing approaches of physical education in colleges and universities, this research employs the classical multivariate statistical technique of factor analysis along with an iterative random forest algorithm. To evaluate the characteristics of machine learning for physical education curriculum, a comprehensive evaluation is performed, and an optimized development strategy is proposed to promote the machine learning-based physical education. Numerous tests were carried out using machine learning classifiers and random forest’s decision tree. The experimental findings reveal that the suggested approach performs better than current models in terms of evaluation’s weight accuracy and calculation time. As a consequence, the suggested model may better match the standards of physical education teaching quality evaluation, i.e., the study's findings are precise and substantial, demonstrating the study's effectiveness.