Abstract
The purpose of this paper is to find the relationship between player data and their salaries through data analysis and modeling of NBA players. Unlike previous papers, this paper divides players into five parts according to different positions and uses more independent variables than previous studies to build models. This article uses 53 sets of player data and their corresponding salaries and conducts preliminary judgment and screening of the data according to the correlation and Boruta formula, and then uses LASSO, RIDGE, Elastic Net, and Random Forest to perform secondary screening and modeling of the data for different positions respectively. The initial screening narrowed the variables from 50 to less than 30. The LASSO and Elastic Net models R-squares are around 0.8, and RIDGE and Random Forest models R-squares are about 0.5. This means that a certain item of LASSO and Elastic Net has a good fit for the data. Then the players income can be predicted according to the two models to help the team build an excellent team with the least amount of money.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
