Predicting Football results with Evolutionary Algorithms

The majority of my work is involved with machine learning using biologically inspired techniques, focusing on classification problems. I run my algorithms on benchmark datasets to test their validity and the effect of various parameters, and then these are used in real life medical applications. Trials can take a long time to prepare, and the data collection process can be somewhat challenging. The group I’m involved with researchs Neurodegenerative Diseases, particularly Parkinson’s Disease. However locating a sufficient number of patients to assess can provide problematic. Ideally machine learning methods are trained on datasets in the hundreds of entries, but measuring hundreds of patients - some with severe movement disabilities - takes an extremely long time.

Thus I’ve been looking for another real world problem to test my methods on. I like my sports, particularly ice hockey and football, and so have previously analysed match results as a way of practicing my web scraping skills. I like to bet a little during major competitions, normally the World Cup, Olympics hockey, Stanley Cup finals and Wimbledon. I wondered if I could use my algorithms to predict match results and help beat the bookies! Also some interesting data analysis could be done, see which league is most predictable and so on.

I’ve set up a system to predict football matches based on previous results. I’m still tweaking to find optimal parameter values, but the aim is to have a single classifier finished for week 5 of the Premier League, which I will assess on each week’s matches. Week 5 because it will use each team’s previous results to predict and so I need to wait for sufficient data to build up. It’s all very well training these algorithms but using them for prediction is where the fun lies. I’ll set up a weekly page here detailing its successes and failures, and also a Twitter feed, so watch this space.

comments powered by Disqus