Multidimensional Regression Analysis and Dynamic Adjustment Modeling of Students’ Physical Fitness Data in Physical Education Teaching in Colleges and Universities

College students' physical fitness is an important part of national health, and analyzing physical fitness data in college physical education teaching helps to dig out the factors affecting students' physical fitness and adjust the teaching plan in time.The article reviews some basic regression tools and selects variables such as BMI dietary habits for logistic regression analysis to analyze the factors affecting students' physical fitness.The similarity, uncertainty and dissimilarity between students and their friends are calculated by Top-N recommendation set algorithm, and the physical education teaching program is dynamically adjusted with the new SFD recommendation algorithm.Finally, values were assigned to different movement banks and risk factors, and the experts' agreement with the new adjusted program was examined.The intensity of physical activity had the greatest relationship with passing or failing physical fitness among all factors (regression coefficient = 0.927, p<0.05), while balanced meat and vegetarian meals helped to enhance students' physical fitness.Most of the professional physical fitness coaches agreed with the teaching adjustment strategies given by this methodology and the level of agreement was high (>70%), reflecting the rationality and feasibility of this study.