篮球
排名(信息检索)
回归分析
秩(图论)
统计
线性回归
计算机科学
差异(会计)
变量
变量(数学)
统计模型
计量经济学
数学
人工智能
经济
地理
考古
数学分析
会计
组合数学
作者
Jeremy Mertz,Laurie Hoover,Jean Marie Burke,David Bellar,Megan L. Jones,Briana Leitzelar,W. Lawrence Judge
标识
DOI:10.1080/24748668.2016.11868925
摘要
The purpose of this investigation was to present a statistical model to rank the top National Basketball Association (NBA) players of all-time. Creating a statistical model to rank players may help sport scientists determine important variables for player analysis, as well as aid coaches in the development of basketball-specific, data-driven performance indicators. Nonetheless, computing this type of model is difficult due to the plethora of individual player statistics and achievements that require consideration, as well as the impact of changes to the game over time on individual player performance analysis. This study used linear regression to create a reliable model for the top 150 player rankings in NBA history. The independent variables within the regression equation included points per game (PPG), rebounds per game (RPG), assists per game (APG), win shares per 48 minutes (WSPER48), and number of NBA championships won (CHMPS). The results revealed that PPG, RPG, APG, and CHMPS were all necessary for an accurate regression model, but WSPER48 was not a statistically significant predictor. The four significant independent variables explained 53% of the variance in player ranking, and further attempts to simplify the regression model were ineffective. The results of the present study also indicated that the commonly-espoused variable WSPER48 did not add statistical merit to the ranking of the all-time greats.
科研通智能强力驱动
Strongly Powered by AbleSci AI