Predicting college basketball match outcomes using machine learning techniques: some results and lessons learned

TL;DR

This paper evaluates various machine learning techniques for predicting college basketball match outcomes, highlighting the importance of features over models and identifying an upper limit to prediction accuracy.

Contribution

It provides an empirical comparison of ML paradigms in sports prediction and offers insights into feature importance and predictive limitations.

Findings

Attributes are more important than models.

There is an upper limit to predictive quality.

Different ML paradigms were evaluated on the task.

Abstract

Most existing work on predicting NCAAB matches has been developed in a statistical context. Trusting the capabilities of ML techniques, particularly classification learners, to uncover the importance of features and learn their relationships, we evaluated a number of different paradigms on this task. In this paper, we summarize our work, pointing out that attributes seem to be more important than models, and that there seems to be an upper limit to predictive quality.