Hosniyeh S Arian - Department of Computer Engineering and Information Technology, Islamic Azad University ,Qazvin
Mohammad J Tarokh - ssociate professor, IT Group - Faculty of Industrial Engineering, K. N. Toosi University of Technology Tehran
Omid R. B Omid R. B - Department of Computer Engineering and Information Technology , Urmia University of Technology Urmia

Influential users who diffuse information and their followers have interest to thisinformation finally they can maximize diffusion in social networks. Influential users havedifferent influence in diversity domain specificity for instance user may have strong influencein a special topic and another topics have weak influence. So a proposed method presentedfor identifying influential users based on domain specificity in this paper. This methodidentified influential users based on domain specificity. When users registered in socialnetworks and initiated to activity in these social networks, all of their actions and cooperationsaved in Data Base. features of user‟s profile and user‟s actions (e.g. retweet) that influenceon diffusion determined by multiple regression and user‟s contents categorized based onkeywords by TF-IDF and finally influential users identified by Tree Regression based ondomain specificity in this paper. The detail of this method discussed the fallowing of paper.In order to evaluate the proposed method on Twitter offer application program interface. 420users selected randomly, they fallow their friends, join to different groups, and generateddiversity tweets on Twitter. The main feature, which distinguishes this method from thepreviously reported methods, is in two key respective. First previous studies have quantifiedinfluence in terms of network metrics for instance number of retweet or page rank, ourproposed method measured influence in terms of the size Tree Regression. Second thefocuses of previous studies were based on the structural of diffusion and feature of contentbut Influential users have different influence in diversity domain specificity so in our proposed method focused on this feature. Results showed that accuracy of proposed method is 0.69.

advertising, content, diffusion, domain specificity, influential users, social networks