Predictive Football Analysis: Leveraging a Random Forest Model and Double Deep Q-Network for Enhanced Performance Analysis

Voiceover

Presenter Information(s)

Anirudh Gadepalli

Abstract or Description

In the fast-paced landscape of sports, technology and algorithms are becoming pivotal forces, creating a new era of performance and allowing athletes to rise to higher levels. Offenses and defenses are constantly evolving and transforming, and with the help of technology, teams are becoming better than ever, and coaching is as difficult as ever.

I present two algorithms to approach coaching sports. In the current football arena, the accurate prediction of the next play has been a longstanding challenge. Therefore, I used an optimized Random Forest Model (RFM) to anticipate what play a team might run, enabling coaches to strategize and teams to create better defenses. Along with that, I developed a second network using Deep Double Q-Learning (DDQN) to attempt to simulate an offense that a coach would call for his players.