首页 | 本学科首页   官方微博 | 高级检索  
     检索      

SOM和PCA对体质健康数据的模式识别及可视化分析——以学生体质地域特征为视角
引用本文:石晓峰,王飞,赵阳.SOM和PCA对体质健康数据的模式识别及可视化分析——以学生体质地域特征为视角[J].天津体育学院学报,2015,30(4):282-287.
作者姓名:石晓峰  王飞  赵阳
作者单位:1.山西大学体育学院,山西太原030006;;2.山西大学体育科学研究所,山西太原030006;2.山西大学体育科学研究所,山西太原030006
基金项目:国家自然科学基金项目(项目编号:41401020);山西省软科学研究计划项目(项目编号:2015041026-4);山西省高等学校创新人才支持计划资助
摘    要:随着国家学生体质健康数据量的剧增,体质健康的大数据分析及可视化成为体质研究的重要内容。自组织特征映射网络(Self-Organizing Map,SOM)方法和主成分分析(Principal Component Analysis,PCA)法对处理高维海量数据具有独特优势及可视化特点,从而成为大数据模式识别和 可视化分析的重要工具。以山西某高校6 531名学生体质健康数据为例,以学生体质地域差异为视角,用SOM方法定性识别了学生体质健康的地域 特征,用可视化PCA方法分析学生体质健康的影响因子及解释因子的地域特征。结论:SOM和PCA方法可用于体质健康数据模式识别和可视化分 析。SOM和PCA的实例分析揭示了学生体质的地域特征,分析显示,体重和BMI 指标具有地域一致性,是影响学生体质健康的最重要因素,也是学 生体质健康现状的主要解释变量;女生体质健康的地域差异相对较大,男生体质健康的地域差异较小;可视化PCA结果还揭示了,学生体质健康指 标的聚类特征也具有地域一致性。文章从实证角度论证了SOM和PCA方法在体质健康数据模式识别和可视化分析中的应用,也为体质类大数据分 析提供了初步思路。

关 键 词:自组织特征映射网络  主成分分析  模式识别  数据可视化  体质健康  地域特征

Pattern Recognition and Visualization of Physical Fitness Data using SOM and PCA: Based on Geographical Features Perspective of Student Fitness
SHI Xiaofeng,WANG Fei and ZHAO Yang.Pattern Recognition and Visualization of Physical Fitness Data using SOM and PCA: Based on Geographical Features Perspective of Student Fitness[J].Journal of Tianjin Institute of Physical Education,2015,30(4):282-287.
Authors:SHI Xiaofeng  WANG Fei and ZHAO Yang
Institution:1.SchoolofPE,ShanxiUniversity,Taiyuan030006,China;;2.SportsScienceResearchInstitute,ShanxiUniversity,Taiyuan 030006,China;2.SportsScienceResearchInstitute,ShanxiUniversity,Taiyuan 030006,China
Abstract:With the data explosion on physical fitness, big data analysis and data visualization have become one of important contents on adolescent physical health research. Data handling techniques of self-organizing map analysis (SOM) and principle component analysis (PCA) are important methods in pattern recognition and data visualization due to their unique features. With the scope of geographical features perspective based on 6 531 college samples, geographical patternofphysicalfitnesswasqualitativelyidentifiedandvisualizedbythemeansofSOM.Meanwhile,thekeyinfluencingfactorsandrelevantexplainingvariables of physical fitness were visualized and extracted to demonstrate the geographical differences by PCA. Results showed that SOM and PCA are powerful tools for pattern recognition and data visualization on physical fitness research. The analysis of the example indicated the geographical consistency of body weight and BMI, whichwerethemostinfluencingfactorsonphysicalfitness,aswellastheexplanatoryvariablesofphysicalfitnessstatus.Therelativelargegeographicaldifferences of physical fitness were observed in female adolescents, but with less geographical differences in male adolescents. The visualized PCA results also revealed the geographical consistency of clustering characteristics on physical fitness variables. The research demonstrated the powerful tools of SOM and PCA on the physical fitnessresearch,andwhichalsoprovidedanewperspectivetostudythephysicalhealth.
Keywords:SOM  PCA  pattern recognition  data visualization  physical health  geographical feature
本文献已被 万方数据 等数据库收录!
点击此处可从《天津体育学院学报》浏览原始摘要信息
点击此处可从《天津体育学院学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号