Analýza reálných dat ze hry League of Legends

The aim of this bachelor thesis is to analyse real data using data mining techniques and interpret the results of the analysis to help the maximum number of players who play the game League of Legends. The analysis precisely follows the CRISP-DM methodology and uses correlation matrices, k-means clustering, and gradient boosted decision trees provided by the RapidMiner tool. The selected dataset contains information about matches of ranked queues in League of Legends and is available for download from the Kaggle.com website for free. The dataset undergoes numerous transformations and is divided into four different subsets: team statistics, player statistics, champion statistics and drafts. The result of the team statistics analysis is ranking the ideal importance of game objectives. The result of the player statistics analysis is ranking the importance of selected game elements for each of the five roles. These results can be used to improve the players' decision-making process in every game. The result of the champion statistics analysis creates a new segmentation of champions, which can be used to expand the existing segmentation and thus make it easier for new players to select champions. The result of the draft analysis mainly confirms the relatively low importance of this game element at the lower to middle level of playing.