A Survey of Forecasting the Salary for Players in National Basketball Association

Abstract

The National Basketball Association is the men's professional basketball league in North America. The National Basketball Association league has some of the highest paid professional athletes in the world. The National Basketball Association is the oldest and most prestigious basketball league in the world, which makes the study of the NBA particularly important. During the 2011-2012 season, the season was cancelled after senior staff and team members failed to reach an agreement due to serious differences of opinion in labour management negotiations. The main reason was that the NBA believed that the players' salaries were too high, which left little profit for the NBA and even increased the NBA's losses. Since the ideal interests of management and the interests of labor did not exactly match, both sides made concessions and reached an agreement. Finally, the season resumed. For most NBA clubs, the salary of each player is usually confirmed before the performance in the next game. However, this makes the question of how to determine the amount more difficult. Players’ salaries should match their performance. If the player does not meet salary expectations, the club is losing money. If a player finds that he gets paid less than his income based on 61 games that year, the club may lose that player and reputation. In this paper, we summarize previous approaches for salary forecasting and categorize them into: (i) linear regression, (ii) modelling approach, and (iii) non-linear regression and (iv) K-nearest neighbour classifier and provide guiding idea for future work.