This repo contains materials for MSCA 31009 Machine Learning and Predictive Analytics project.
Professor: Dr. Arnab Bose
Team:
Adrienne Wang [email protected]
Serena Hahwen Xu [email protected]
Michael Thompson [email protected]
Oleksiy Anokhin [email protected]
In this project we built several machine learning and deep learning models to predict individual MLB player performance per game based on FanDuel’s scoring system (please see mode about project data in our final presentation).
Professors at MIT made $15,000 on Daily Fantasy NHL over an 11-week period using a combination of RotoGrinders and Daily Fantasy Nerd predictions (63% and 28%, respectively) and constructing lineups using linear programming. In addition, they published a paper about their work - Picking Winners Using Integer Programming
They mention they performed a similar (but only hypothetical) study on baseball that also proved successful.
Therefore, we have used predictions from the website RotoGrinders as the benchmark to evaluate our predictions.
Our final models outpeformed the established benchmark (please see the additional information in our presentation).