Document Type
Thesis
First Faculty Advisor
Rick Gorvett
Second Faculty Advisor
Jim Bishop
Keywords
machine learning; statistics; analytics
Publisher
Bryant University
Rights Management
CC-BY
Abstract
Throughout the history of the NBA, decisions regarding the signing of free agents have been riddled with complexity. Franchises are tasked with finding out what players will serve as optimal free agent signings prior to seeing them perform within the framework of their team. This study hypothesizes that the adequacy of an NBA free agent signing can be modeled and predicted through the implementation of a machine learning model. The model will learn the necessary information using training and testing data sets that include various player biometrics, game statistics, and financial information. The application of this machine learning model will determine the characteristics of what the ideal NBA free agent signing looks like. This data modeling is of interest for the entirety of the NBA as it can determine appropriate contract values based off players’ careers up until any given point in time. The overall goal of this research is to aid NBA franchises in efficiently signing players.
