Date on Honors Thesis
Spring 4-29-2026
Major
Computer Information Systems
Examining Committee Member
Dr. Cemil Kuzey, Advisor
Examining Committee Member
Dr. Victor Raj, Committee Member
Examining Committee Member
Dr. Hassan Mistareehi, Committee Member
Abstract/Description
Artificial Intelligence (AI) and Machine Learning (ML) are more commonly becoming tools that organizations use to prevent, detect, and deter fraud. Like other organizations, the amount of fraud that the National Collegiate Athletic Association (NCAA) faces is growing at an alarming rate. The primary issue involves athletes that undermine the integrity of their sport by purposefully playing below their true potential as participants in broader betting-related schemes. This thesis focused on Division 1 Men’s Basketball evaluates the viability of using supervised ML to detect anomalies in team performance over games in recent seasons. Anomalies, like red flags in the context of fraud, will indicate that performance manipulation was more likely in that game. The purpose is to transform game statistics into actionable information that increases oversight and monitoring. Training and testing the model with games from the 2023-24 and 2024-25 seasons, our random forest classifier achieved an average accuracy of 45%. Although being the highest in modern literature, we conclude that this approach cannot viably be applied as a point shaving detection tool with the current feature set.
Recommended Citation
Whitaker, Eli G., "Investigating Machine Learning as an Anomaly Detection Tool in College Basketball" (2026). Honors College Theses. 314.
https://digitalcommons.murraystate.edu/honorstheses/314
Included in
Business Analytics Commons, Business Intelligence Commons, Management Information Systems Commons, Management Sciences and Quantitative Methods Commons
Additional Author Comments
For publication reasons, the full document will not be uploaded to the Digital Commons. Full document can be made available to individuals upon request.