| Abstract [eng] |
In this work there is predicting basketball game winner be using data from the strongest European basketball league Euroleague and VTB basketball league protocols. There were used stepwise logistic regression methods and neural networks for predicting results. Also, based on forecast results there were found the probabilities to win a game for the team, which were used for betting in the betting firms. The aim of the work is to find out which mathematical methods a better for forecasting: logistic regression or neural networks. Also to find out is it possible to make money from betting companies, if you know you own probabilities of winning. The result is that better forecasting is from neural networks, but more money can be made using logistic regressions. In the future, working with more data it is possible to get better rasults. |
