Abstract—Making predictions for the sport competitions, which are followed by wide masses, has always been an interesting field for sport fans, bettors, researchers and etc. despite the complexity and uncertainty of many factors. The result of a sport competition is affected by many independent variables and factors. The number of the variables that are included in the calculation affects the accuracy of the prediction. It is difficult for an ordinary punter to cope with these factors that are in high numbers and that have a high complexity. On the other hand, a bookmaker must consider all the factors that might affect the result. When a bookmaker determines the odds with some delicate calculations, it is actually digitizing all the above-mentioned complex factors. In this way, the consistency of the betting odds of past competitions becomes a good indicator to be able to make predictions. In this study, a prediction model has been suggested for football game, which is more common than the other sports branches. In this model, the basic design approach is to measure the similarity between competitions in a way based on betting odds. The model was enhanced with the performance data obtained by the past games. Adding the risk analysis option to the model decreased the margin of error in games predicted at a great deal. The Super League of Turkey competitions were used to test the model in which the k-Nearest Neighbor Algorithm was preferred as the estimation technique.
Index Terms—Bookmaker odds, estimation, football, k-nearest neighbor algorithm.
The authors are with the Selcuk University, Engineering Faculty, Department of Computer Engineering, Konya, Turkey (e-mail: email@example.com, firstname.lastname@example.org).
Cite: Engin Esme and Mustafa Servet Kiran, "Prediction of Football Match Outcomes Based On Bookmaker Odds by Using k-Nearest Neighbor Algorithm," International Journal of Machine Learning and Computing vol. 8, no. 1, pp. 26-32, 2018.