Abstract

Each year, sixty-four teams receive invitations to the Women's Division I National Championship. These sixty-four teams are split into two thirty-two team categories known as at-large and automatic bids. Every year, analysts attempt to model and predict how the committee decides on which teams receive at-large bids into the tournament and the seeding of the sixty-four chosen teams. Methods of logistic regression and random forests were used to model the committee's selection process. Multiple statistics were gathered to be used as predictors within the models created. These methods were tested on data from the 2013-2019 seasons. Overall, the random forest method was found to be the most predictive in modeling the at-large selection process and seeding.

Advisor

Pasteur, Drew

Department

Mathematics

Disciplines

Applied Mathematics

Publication Date

2019

Degree Granted

Bachelor of Arts

Document Type

Senior Independent Study Thesis

Share

COinS
 

© Copyright 2019 Hunter D. Coia