基于因子分析和K-Means算法对NBA得分后卫的功能性分析

Functional Analysis of NBA Shooting Guard Based on Factor Analysis and K-Means Algorithm

以2012—2013赛季至2018—2019赛季美国职业篮球联赛得分后卫的基础数据和进阶数据为基础,运用因子分析和K-Means聚类算法建立聚类模型,对球员类型进行分析。相比于传统方法利用基础数据对所有球员进行综合评比,笔者新增了进阶数据并通过机器学习方法反映球员在场上的真实定位和功能。结果表明:得分后卫可以从领袖因子、进攻贡献因子、防守贡献因子、三分球能力、投篮稳定性以及传球能力六个指标,分为绝对核心、角色球员、外线核心、顶级侧翼防守者、外线球员、组织核心、“3D”球员、华而不实型八类球员。进阶数据的使用更能体现球员的真实价值。建议:中国职业篮球联赛(CBA)应顺应大数据时代的潮流,引进先进的数据统计设备,健全数据统计类型,注重进阶数据的统计,培养数据分析人员,通过数据分析球员的功能和类型,为球队组建、教练用人和人才培养提供科学的指导。

Based on the basic and advanced data of NBA shooting guards from 2012—2013 season to 2018—2019 season,this paper establishes a clustering model by using factor analysis and K-Means clustering algorithm to analyze the players' types.Compared with the traditional methods of using basic data to comprehensively evaluate all players,this paper adds advanced data and reflects players' real positioning and function by machine learning method.It shows that shooting guards can be divided into eight classes,including absolute core players,role players,core perimeter players,top wing defenders,perimeter players,playmakers,3D players and pompous players from six factors,namely leader factor,offensive contribution factor,defensive contribution factor,three-point shooting ability factor,shooting stability factor and passing ability factor.The usage of advanced data can reflect the value of players better.It is suggested that China Basketball Association(CBA)conform to the trend of the era of big data,introduce advanced statistical equipment,perfect the data statistic type,pay attention to statistics of advanced data,train data analysts,analyze the function and type of players based on data in order to provide scientific guide for team formation,coaches employment and talent cultivation.