NBA Player Salary Analysis based on Multivariate Regression Analysis

Abstract

The National Basketball Association (NBA) is one of the four major professional sports leagues in the United States and holds substantial economic value. Since the creation of the NBA, the issue of NBA player salaries has become increasingly relevant and the disproportionate salaries of players has emerged as a challenge that obstructs the growth of the league. This paper will employ data from 331 NBA players over the 2020-2021 season to develop multiple linear regression models for analyzing appropriate player salaries. The multiple regression analysis was conducted using basic and advanced data of players as independent variables, which include dozens of variables such as player's age, player's number of games, and player's average score, and player salary as dependent variable. In order to analyze the results more accurately, several types of multiple linear regression models are used in this paper: Ordinary least squares, Ridge regression, Lasso regression, Elastic net regression. The outcome of the regression model for NBA player salary will contribute to enhancing the commercial worth of the NBA and provide constructive input to the league and the team. This research found investors place higher significance on players' three-point shooting proficiency, defensive abilities, and teamwork capacity, which aligns with the current playing style of the NBA.