About ML-MMA
This project showcases fight predictions generated using Autogluon's ensemble of machine learning models. All machine learning models have been trained with Autogluon using an 80, 10, 10 split (train, val, test). Only UFC fights and UFC fighters who have more than one fight in the UFC can be predicted. Women fights have been removed from my dataset and are not predictions I make due to the fight outcome unpredictiability and lack of historical accurate data among all their weightclasses.
About Me: I am a student at The University of Iowa studying business analytics and information systems. The UFC has been a sport I have followed since as long as I can remember with my earliest memories being Anderson Silva defending his title against Chael Sonnen. As I got older I became interested in betting on the fights I was watching, and I felt with my knowledge of the sport I could make solid predicitons. As I began to study BAIS I came to realize while I may be able to make good predictions myself from years of watching the sport, the hard data will always be able to do it better, and ML - MMA is the result.