An Accurate Assessment Model of Big Data-Based Physical Education Programs in Colleges and Universities for Student Health Enhancement

Abstract

In this paper, we design a research framework for college student health level data that integrates K-means clustering method and Apriori algorithm to establish an assessment model between physical education courses and student health. The maximum minimum distance and inter-sample density are used to improve the sample point categorization of K-means clustering and select more appropriate initial clustering centers. Apriori, a linkage linkage profiling class algorithm, is used to deeply mine frequent item sets at multiple levels to improve modeling efficiency and accuracy. It was found that five strong association rules were mined for each of male and female college students, with support and confidence ranges of 9.214%-10.925%, 57.071%-90.426% (male) and 11.351%-15.713%, 55.423%-67.153% (female), respectively. Male and female college students were clustered into 5 categories according to their physical fitness test scores respectively. Physical fitness level = 72.735 + 0.689*exercise intensity + 0.527*exercise time + 0.641*exercise frequency.