Algorithm of the Football Match Result Prediction on the Basis of Neural Networks

The authors present the research of using one of the approaches of machine learning called neural network in predicting sporting events results. At the moment, the field of machine learning is at the peak of interest among the general public and is used in solving a large aspect of the problems of the modern world. One of the main reasons for neural networks and machine learning to be very popular today, is the availability of computer resources for everyone, which gave the opportunity to use them for almost everybody who needs to solve their problems, not spending a huge amount of money on the resources for neural networks specifically. Among the main advantages and at the same time the reason, that has become the main one in choosing the method of solving the posed problem is the possibility of using neural networks to solve problems in which there is no clear algorithm provided. Another area that has been in the lead for several years now is the bookmaking industry, which popularity is only growing. This is especially evident concerning the most popular kind of sport in the world – football. This situation was partly caused because of computerization which today allows almost every computer user make a bet remotely any time they wish, even during the sport match. Based on the collected data about architecture and methods of using various types of neural networks, as well as in general methods that can be used to solve prediction problems, research was conducted and prediction algorithm was developed. The authors have explained the lack of similar tools in the open access to predict the results of sporting events; the evaluation of the algorithm was carried out by comparing accuracy with other forecasting algorithms that do not involve the use of neural networks. Such testing was carried out on the basis of a publicly available data set obtained from open sources.