Hongxu Ma MSc Dissertation 2014/15

Machine Learning for Modelling Formula One Races

Supervised by N.Lawrence

Abstract

Formula One race is a complex sport which requires good strategies to win. One special feature of this sport is that it can generate a lot of data from both cars and players. This data is interesting since it contains a lot of information which may be useful for team to improve their strategies. A fitting analysis on the data may even potentially make the difference between winning and losing the match.

This project focuses on the practice and qualifying session in Formula One race. The aim is to predict the cut off time for passing qualifying session under the GPy framework. If this predict is applicable it may give teamsome advantages when making strategies. Previous work shows that Gaussian process can be a practical model on this problem. It it powerful, flexible and easy to implement under GPy framework.