Year of Graduation
Matchmaking in the Social Network Game
This paper is devoted to improving the user experience in social casual poker by developing adaptive matchmaking, which considers the player's game behaviour. The general purpose is increase the number of games played by user with automatic table selection. In this paper, considered the analysis of real game dataset, various methods clustering (k-means and spectral) are used for defining user groups with different statistical tests for significance detection and, finally, development of prediction model of played time between different clusters. Also, a separate part of the work is devoted to the analysis of games and psychology of poker. As a result, developed an algorithm that can increase the retention and, therefore, the user lifetime value.