Abstract

To address the question of whether the quality evaluation of physical education teaching in colleges and universities can be analyzed mathematically, this paper proposes an evaluation model of AHP-BPNN based on hierarchical analysis and BP neural network technology. Among them, hierarchical analysis is used to solve the problem of determining evaluation factors, selecting the evaluation factors contained in the index system, preserving those factors that are more important for the teacher evaluation model and removing those that have less influence on the evaluation results. Finally, the BP neural network algorithm is used to establish the teaching evaluation model to solve the problem of artificially assigned index weights, and also to solve the model problem of non-linear relationship between each evaluation index and teaching effect. The accuracy analysis of the model obtained that the average accuracy of the model of this paper on the evaluation factors of physical education teaching quality is 88.83%, and the accuracy of POE evaluation method, fuzzy evaluation method and gray correlation analysis method are 74.33%, 88.83% and 82.83% respectively. In the weight calculation and consistency test of the evaluation index system, the highest weight coefficient of teaching ability was 0.3367, the second highest weight coefficient of teaching attitude and the lowest weight coefficient of teaching effect were obtained. Therefore, the accuracy of the model in this paper is good, and the evaluation system established by the model concludes that schools should pay more attention to the teaching ability of teachers.