Using multiobjective genetic algorithm for the community discovery in social networks

Social network analysis (SNA) is a topic of interest gaining more attention since 2000 years. SNA is in the scope of many different branches of science such as social science, computer science, and biology etc. SAA takes an interest in better understanding complex network by handling relations among animate and inanimate entities as information network. Link prediction among actors in the network, pattern discovery on users? behaviors and community discovery are some study topics in SAA. In this study a multi-objective genetic algorithm is proposed for community discovery in social networks. Multi-objective optimization approaches give better results on problems having multiple criteria. The problem of community discovery in social networks can be considered as a multi-objective optimization problem. Communities are discovered in social networks by using fitness functions in the literature with proposed effective selection approach. Many of the similar studies use classic methods when creating initial population and mutation step which is used for creating new individuals. In our study these processes are made with respect to method which is suitable community discovery problem structure. The proposed method is tested on synthetic and real networks, it is observed that better results are obtained by compared to a similar study. Keywords: Complex networks, community discovery, multi-objective evolutionary algorithm.