基于集成学习模型的大学生体测成绩预测  

Prediction of Physical Performance of College Studentsbased on Ensemble Learning Model

作　　者：黄蓉 韩哲哲 HUANG Rong;HAN Zhe-zhe(Information construction and Management Office,Nanjing University of Posts and Telecommunications,Nanjing 210042,China;School of Information and Communication Engineering,Nanjing Institute of Technology,Nanjing 211167,China)

机构地区：[1]南京邮电大学信息化建设与管理办公室,江苏南京210042 [2]南京工程学院信息与通信工程学院,江苏南京211167

出　　处：《福建体育科技》2023年第6期112-118,共7页Fujian Sports Science and Technology

基　　金：南京工程学院高等教育研究课题(2022ZC09)。

摘　　要：大学生体测成绩是衡量身体素质和健康状况的重要指标,关系到日常生活与学习。通过体测成绩预测,有助于及早发现潜在健康问题,并提供个性化锻炼建议。传统体测成绩预测方法主要基于主观经验和统计策略,存在较大不确定性。为了弥补现有技术不足,本研究提供了一种基于集成学习模型的体测成绩预测新方法。在该模型中,体质特征(如年龄、身高、体重和肺活量)作为输入,极限学习机和决策树作为预测引擎,支持向量回归作为堆栈模型。在模型训练过程中,通过交叉验证方法对超参数进行优化,从而获得最佳预测性能。实践测试表明,集成学习模型能够准确预测体测成绩,其性能明显优于单一预测引擎,展现出良好应用前景。The physical performance of college students is an important indicator to evaluate physical fitness and health status,directly related to daily life and study.By predicting physical performance,it is helpful to detect potential health problems early and provide personalized exercise suggestions.Traditional prediction methods for physical performance primarily rely on subjective experience and statistical parameters,encountering significant uncertainty.To address the shortcomings of the existing techniques,a new prediction method for physical performance based on an ensemble learning model is proposed in this study.In this model,the physical characteristics(such as age,height,weight and vital capacity)are used as inputs,the extreme learning machine and decision tree are applied as prediction engine,and the support vector regression is utilized as the stack model.During the training process,the hyper-parameters are optimized through cross-validation to achieve optimal performance.Practical testing demonstrates that the ensemble learning model can accurately predict performance,outperforming individual prediction engines and presenting promising application prospects.

关 键 词：体质特征 人工智能 堆栈模型 集成学习 成绩预测 

分 类 号：G807.4[文化科学—体育训练]