A substantial amount of research has been on finding models that are used for predicting sports outcomes. More research has been conducted involving other sports, including NCAA basketball, than with the NBA. The focus of my study was to see how accurate artificial neural networks would be in predicting the outcome of NBA games and the margin at which the team would win by. Then comparing those results from artificial neural networks to linear models that see which model would be more accurate. Data from 2008-09 and 2009-10 seasons were used in the models and tested on the 2010-11 season. Another model used was choosing the home team to win and see how accurate this pick would be. Comparing seven different models to find which statistics best fit in predicting NBA outcomes.


Pasteur, Drew




Applied Mathematics

Publication Date


Degree Granted

Bachelor of Arts

Document Type

Senior Independent Study Thesis



© Copyright 2012 Keshia Butler