Seyed Ahmad Moosavi - Dept. of Computer Engineering Imam Reza International University Mashhad, Iran
Mehrdad Jalali - Dept. of Computer Engineering Islamic Azad University Mashhad branch

recently, the online social networks provide a rich resource of Heterogeneous data which its analysis can lead to discover unknown information and relations withinsuch networks. In analysis of social networks data, a challenging issue is the discovery of community including similar nodes and it has widely been studied in the socialnetworking community in the context of the structure of the underlying graphs. The online social networks, additionallyhaving graph structures, include effective information of users within networks, which using this information can lead to improve the quality of communities' discovery. In thispaper, instead of using centrality measures in social networks analysis, to discover leaders and similar nodes,user actions are used and by using these leaders, communities are identified. First, based on Interests and activities of users in networks, we discover some smallcommunities of similar users, and then by using social relations, extend (those) communities. Finally, by conductingdoing empirical studies, the efficiency of our approach on community discovery within the online social networks will be demonstrated.

Online Social Network, Identify leader,Community Detection, Frequent Pattern Mining Algorithm