Skip to content

Deep Learning and Transfer Learning Architectures for English Premier League Player Performance Forecasting: CS229 Final Project

License

Notifications You must be signed in to change notification settings

danielfrees/mlpremier

Repository files navigation

mlpremier

Modeling Player Performance in the EPL via time-series prediction of Fantasy Premier League Points using traditional ML methods, GPT transfer learning, and 1D CNN.

Usage

If you find this useful, please cite:

@article{frees2024deep,
  title={Deep Learning and Transfer Learning Architectures for English Premier League Player Performance Forecasting},
  author={Frees, Daniel and Ravella, Pranav and Zhang, Charlie},
  journal={arXiv preprint arXiv:2405.02412},
  year={2024},
  note={10 pages},
  url={https://doi.org/10.48550/arXiv.2405.02412},
  eprint={2405.02412},
  archivePrefix={arXiv},
  primaryClass={cs.LG},
  month={May 3}
}

Abstract

Here we present a groundbreaking model for forecasting English Premier League (EPL) player performance using convolutional neural networks (CNNs). We evaluate Ridge regression, Light-GBM and CNNs on the task of predicting upcoming player FPL score based on historical FPL data over the previous weeks. Our baseline models, Ridge regression and LightGBM, achieve solid performance and emphasize the importance of recent FPL points, influence, creativity, threat, and playtime in predicting EPL player performances. Our optimal CNN architecture achieves better performance with fewer input features and even outperforms the best previous EPL player performance forecasting models in the literature. The optimal CNN architecture also achieves very strong Spearman correlation with player rankings, indicating its strong implications for supporting the development of FPL artificial intelligence (AI) Agents and providing analysis for FPL managers. We additionally perform transfer learning experiments on soccer news data collected from The Guardian, for the same task of predicting upcoming player score, but do not identify a strong predictive signal in natural language news texts, achieving worse performance compared to both the CNN and baseline models. Overall, our CNN-based approach marks a significant advancement in EPL player performance forecasting and lays the foundation for transfer learning to other EPL prediction tasks such as win-loss odds for sports betting and the development of cutting-edge FPL AI Agents.

Data scraped from vaastav's FPL Data Repository

About

Deep Learning and Transfer Learning Architectures for English Premier League Player Performance Forecasting: CS229 Final Project

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published