An evolutionary game competitive model for increasing accuracy and maximizing influence in social networks

Given the importance of maximizing influence in a social network, studies in this field often seek to find the nodes that have the most influence on the social network if designated as primary seeds. In this study, to reduce the complexity of computation algorithms, the problem is divided into several groups that aim to find a group of influential people among users of a social network. In this paper, a framework is introduced for solving the problem of influence maximization, which is based on the member clustering by the K means method, to improve the classification of network users, the data are weighted and the problem is modeled and analyzed as an evolutionary game. Finally, calculate its evolutionary stable strategy. This framework has been tested on real social network data for Abrar University students and we have achieved results such as increasing classification accuracy, reducing error function and finding a stable strategy in the community.