Victory prediction in League of Legends using Feature Selection and Ensemble methods

Prediction of winners in the online video games has become an important application for machine learning based prediction models. The main goal of the present study is to achieve a good prediction rate for a popular Electronic sport called League of Legends. League of Legends is a Multiplayer Online Battle Arena (MOBA) game that combines intensity of a Real-time strategy with various Role-playing elements. Feature selection is done and only relevant features that affect the match outcomes are considered. Prediction is done by using ensemble models of classification algorithms and the performance was evaluated. The important performance metrics and their influence on each game model were also analyzed. The results show that the reliable match result prediction is possible in the League of Legends game.