基于机器学习算法的运动员训练效果评估研究  

Research on athlete training effect evaluation based on machine learning algorithm

作　　者：岳志强 YUE Zhiqiang(Shanghai Jianqiao University,Shanghai 201306,China)

机构地区：[1]上海建桥学院,上海201306

出　　处：《电子设计工程》2021年第20期110-114,共5页Electronic Design Engineering

摘　　要：为了降低运动员训练效果评估误差,便于教练员合理制定训练计划,提出了基于机器学习算法的运动员训练效果评估方法。选取心率、摄氧量、血红蛋白等生理指标作为支持向量机输入向量,将运动员训练效果评估期望值作为输出建立训练样本,设置支持向量机参数,以评估误差最小为学习目标。支持向量机对训练样本进行学习建立运动员训练效果评估模型,选取某大学运动员作为研究对象,通过足球、篮球、排球、游泳、跑步5项运动进行训练效果评估测试,结果表明,所提方法的评估运动员训练评估误差低,具有较高的实际应用价值。In order to reduce the evaluation error of athletes'training effect and make it convenient for coaches to make training plans reasonably,the evaluation method of athletes'training effect based on machine learning algorithm is proposed.The physiological indexes such as heart rate,oxygen uptake and hemoglobin are selected as the input vectors of support vector machine,and the expected value of training effect evaluation of athletes is taken as the output to establish training samples.The parameters of support vector machine are set up.With the minimum evaluation error as the learning goal,the training samples are learned by support vector machine,and the training effect evaluation model is established.A university sports event is selected mobilization as the research object,through football,basketball,volleyball,swimming,running five sports training effect evaluation test,the results show that the evaluation of athletes training evaluation error is low,has high practical value.

关 键 词：机器学习算法 训练效果 支持向量机 高维特征空间 指标矩阵 

分 类 号：TN01[电子电信—物理电子学]