Learning to predict soccer results from relational data with gradient boosted trees
- 221 Downloads
We describe our winning solution to the 2017’s Soccer Prediction Challenge organized in conjunction with the MLJ’s special issue on Machine Learning for Soccer. The goal of the challenge was to predict outcomes of future matches within a selected time-frame from different leagues over the world. A dataset of over 200,000 past match outcomes was provided to the contestants. We experimented with both relational and feature-based methods to learn predictive models from the provided data. We employed relevant latent variables computable from the data, namely so called pi-ratings and also a rating based on the PageRank method. A method based on manually constructed features and the gradient boosted tree algorithm performed best on both the validation set and the challenge test set. We also discuss the validity of the assumption that probability predictions on the three ordinal match outcomes should be monotone, underlying the RPS measure of prediction quality.
KeywordsPrediction challenge Relational data Soccer Gradient boosted trees Relational dependency networks Sports Forecasting
The authors acknowledge support by project no. 17-26999S granted by the Czech Science Foundation. Computational resources were provided by the CESNET LM2015042 and the CERIT Scientific Cloud LM2015085, provided under the programme “Projects of Large Research, Development, and Innovations Infrastructures”. We thank the anonymous reviewers for their constructive comments.
OH processed the data, developed the concept, ran and evaluated the experiments and participated in the writing of the manuscript, GŠ provided consultations through each stage of the work and participated in the writing of the manuscript, FŽ supervised the work and wrote the final manuscript.
- Chen ,T. & Guestrin, C. (2016). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (pp 785–794). ACM.Google Scholar
- Lazova, V. & Basnarkov, L. (2015). PageRank approach to ranking national football teams. arXiv preprint arXiv:1503.01331.
- Natarajan, S., Khot, T., Kersting, K., Gutmann, B., & Shavlik, J. (2010). Boosting relational dependency networks. In Online Proceedings of the international conference on inductive logic programming, 2010 (pp. 1–8).Google Scholar
- Odom, P. & Natarajan, S. (2016). Actively interacting with experts: A probabilistic logic approach. In Joint European conference on machine learning and knowledge discovery in databases (pp. 527–542). Springer.Google Scholar
- Pollard, R., & Pollard, G. (2005). Home advantage in soccer: A review of its existence and causes. International Journal of Soccer and Science, 3(1), 28–44.Google Scholar
- Van Haaren, J. & Davis, J. (2015). Predicting the final league tables of domestic football leagues. In Proceedings of the 5th international conference on mathematics in sport (pp. 202–207).Google Scholar
- Van Haaren, J. & Van den Broeck, G. (2015). Relational learning for football-related predictions. In Latest advances in inductive logic programming, world scientific (pp. 237–244).Google Scholar