Titulo Estágio
Predicting Football Events using Data Mining Techniques
Áreas de especialidade
Sistemas Inteligentes
Engenharia de Software
Local do Estágio
University of Coimbra -- Department of Informatics Engineering
Enquadramento
Today, football is one of the most played sports worldwide and because of that, it has seen a great evolution. There are many kinds of agents involved in the game and the balance between the teams is growing. Because of that, predicting a specific event of the game (before or during the game) plays an important role. However, this task is increasingly difficult to achieve.
Objetivo
Using real football data enclosing different professional leagues (e.g. La Liga, Premier League), the goal of this project is to predict game events using data mining techniques. The first event will consist on the prediction of final game result but other events can be predicted e.g number of goals scored in each part.
The data collection will be performed using information provided by sport websites like zerozero, sapodesporto and the validation of the results provided by the data mining techniques will be validated using information from sports betting website. At the end, the results should highlight some important aspects like: competitiveness of the football league, the set of variables that can better predict a specific soccer event attending to the professional league, the best algorithm which adapts to this environment with a strong stochastic component among others.
Plano de Trabalhos - Semestre 1
-Study and analysis the state of the art concerning the prediction of football events;
-Data gathering of football data from sports website.
Plano de Trabalhos - Semestre 2
-Implementation of different data mining techniques;
-Evaluation of the implemented techniques using the bet sites;
-Comparison of obtained results and conclusions;
-Writing the report of the Master Degree Thesis.
Condições
The student will work with real football data
Orientador
Pedro Henriques Abreu
pha@dei.uc.pt 📩