2017 Student Research Conference:
30th Annual Student Research Conference

Analysis of Machine Learning Algorithms to Predict March Madness Results


Monica Singh* and Paulie Massey
Dr. Ruthie Halma, Faculty Mentor

Each March, millions of Americans compete to predict a perfect bracket for the men’s NCAA basketball tournament. Although many guessers come close, there is no known formula which guarantees a perfect bracket. Attempts to predict the bracket by using machine learning algorithms have yielded a log loss (also known as predictive binomial deviance) of 0.48 at best. This project aims to determine the best classification algorithm to predict wins and losses in the 2017 March Madness tournament, using Ensemble algorithms.  A high-performing classification model proves that sports predictions can be accurately determined by machine learning techniques.

Keywords: machine learning algorithms, data mining

Topic(s):Computer Science

Presentation Type: Oral Paper

Session: 312-2
Location: VH 1320
Time: 1:15

Add to Custom Schedule

* Indicates the Student Presenter
Contact SRC Webmaster  |  SRC Privacy Policy