自拍麻豆另类

MSDS Students at Scholars Week - College Football Recruiting Using Machine Learning Techniques

Colette Barca, Keith Osani, Nisha Srishan, and William 鈥淏rady鈥� Wulster are inaugural members of the Master鈥檚 in Data Science program. During their MATH 570 Applied Statistics course with Professor Osei Tweneboah, they designed and completed a research project to predict high school football recruits鈥� college commitments using several Machine Learning techniques.

The four designed their project with the understanding that college football is big business. To create a winning team that will continue to generate revenue, Division I Football Bowl Subdivision (FBS) Power Five schools need to choose their recruits wisely. Recruits are selected
based on certain player attributes. Power Five schools are the most frequented schools by NFL scouts, making these schools highly desired by recruits. In this regard, Barca, Osani, Srishan, and Wulster built a model to predict whether a high school recruit will commit to a school in one of the Power Five Conferences. Such a model could allow a school to optimize its recruiting process, maximizing its return on investment. They used their dataset to fit several Machine Learning models. After completing the project, it was determined the Random Forest model produced the most accurate results. This model also revealed which particular attributes are indisputably the most important predictors of commitment to a Power Five school. This final model has the potential to be successfully used to improve the recruiting process.

Abstract

Poster Presentation

Categories: Data Science, MSDS