Identification of communities in social networks based on game theory with stable coalitions

Due to the availability of more data and the increase of interactive activities in social media, the identification of overlapping associations has been considered. In this paper, a game theory-based approach to identify overlapping associations is proposed. In this method, association detection is modeled as a coalition formation game. In this game, individuals in a social network are modeled as rational actors whose goal is to improve the group's utility, which is achieved by cooperating with other players and forming coalitions. Each player can join multiple alliances, and alliances with fewer players can merge into a larger alliance as long as the joining operation is conducive to the alliance's goals. Therefore, overlapping associations can be identified simultaneously. In this article, two types of methods based on cooperative and non-cooperative game theory have been discussed. The results report is analyzed based on the comparison of association methods in the form of a diagram. It can be seen that the game group and the COFOGA method perform better association.