Exploration of various approaches to modelling football matches.
Will try the following modelling approaches:
- Poisson regression model
- Bayesian Possion regression model
- Bayesian hierarchical model
- Bayesian hierarchical Poisson model
Football predictions are hard and typically, bookies do a better job at it (probably because they hire the most talented statisticians and modellers). This project is not an attempt to accurately predict football matches per se; it is a personal odessey to explore existing approaches.
Additional resources to consider:
- A simple Bayesian procedure for forecasting the outcomes of UEFA Champions League matches
- Dolores: a model that predicts football match outcomes from all over the world
- Soccer predictions using a Bayesian multilevel model
- Bayesian approach to ranking English Premier League Teams
- Predicting Football Matches Results using Bayesian Networks for English Premier League (EPL)
- Bayesian Approach to Classification of Football Match Outcome