基于贝叶斯分位数回归研究大学生体测指标  

Research on Physical Fitness Indicators of College Students Based on Bayesian Quantile Regression

作　　者：陈润玉 虞克明 CHEN Runyu;YU Keming(School of Mathematics and Physics,Anqing Normal University,246003,Anqing,Anhui,China)

机构地区：[1]安庆师范大学数理学院,安徽安庆246003

出　　处：《淮北师范大学学报（自然科学版）》2024年第2期8-14,共7页Journal of Huaibei Normal University：Natural Sciences

基　　金：国家社会科学基金项目(21BTJ040)。

摘　　要：为深入了解大学生身体素质现状及体测指标之间的相关性,采用贝叶斯分位数回归法,对某高校2020级学生的体质健康测试数据进行研究。首先对体质测试指标进行相关性检验和显著性检验,然后针对相关性较强的指标数据进行二次分位数拟合分析。女生的肺活量与身高、50 m短跑与立定跳远、男生的50 m短跑与1000 m长跑、50 m短跑与立定跳远以及1000 m长跑与体测总分之间的关系拟合效果较好,当分位点等于0.25时,模型精度最高。实验表明,贝叶斯分位数回归能全面反映大学生群体体测特征,具有较高的检测精度和良好的稳健性。To comprehensively investigate and comprehend the current state of physical fitness among college students,as well as the interplay between various physical indicators.A pioneering application of Bayesian quantile regression is introduved focused on the physical fitness index of university students in the year 2020.Firstly,correlation tests and significance assessments are conducted on these physical indicators.Sec⁃ondly,quadratic Bayesian quantile fitting is applied to data exhibiting robust correlations.The findings re⁃veal that the proposed quantile regression models offer superior fits for relationships,such as vital capacity and stature,50 m sprint and standing long jump for females,50 m sprint and Men′s 1000 m,50 m sprint and standing long jump for males,as well as Men′s 1000 m and the overall score.And model accuracy is best when quantile=0.25.The results demonstrate that Bayesian quantile regression can reflect the physical characteristics of college studentscomprehensively,and has higher detection accuracy and better robustness.

关 键 词：贝叶斯分位数回归 体育统计 体质健康 大学生 

分 类 号：O29[理学—应用数学]