Research Of Dynamics Users’ Influence In Social Network Based On Users’ Behavior Model

With the rapid development of the Internet, it has become the main way that can send and receive information. Social network is an important field in the Internet today. With the development of the mobile Internet, people can public and share messages in this platform anywhere. Social network is playing a very important role in the message communication platform. As a new power on the Internet, micro-blogging has high speed over other media. It has a lot of knowledge to read. So it has wide influence on finding the information. We can get a lot of real time message through micro-blog, therefore it is widely applied to find the influence of users in micro-blog users. But the current researches on users’ influence exist a phenomenon of unilateral factors, the results of the evaluation does not have the features of accuracy, and those algorithms considering the factors over the whole are complexity.In view of the above problems, this paper presents an dynamic algorithm(BDR algorithm) based on users’ behavior to evaluate the user influence. Firstly research the method of user evaluating problem, and summarizes the advantages and disadvantages of these methods, found the forwarding behavior and time factors affect the user’s influence. In this paper, micro-blog user forwarding behavior in the time interval obeys the power-law distribution from human behavior dynamics and user behavior analysis, so this article makes use of the forwarding relationship building network diagram, and according to the user’ forwarding situation it celebrates user relationships into frequent forwarding relation and occasional forwarding relation, and every once in a while the two relationships update, so it can realize the dynamic relationship chart updating. Dynamic forwarding diagram provides the basis for the study of user influence condition evaluation. Then, according to the Page Rank algorithm and Twitter Rank algorithm idea, this paper presents a dynamic model--BDR algorithm to evaluate users’ influence. This algorithm bases on the relational network dividing, different network can struct different forwarding probability model for the relationship, and then integrated the two forwarding probability model to get the distribution weight in general, finally combining the Page Rank algorithm it realizes dynamic node influence algorithm based on users forwarding behavior. Finally, through the experiment statistical analysis on the BDR algorithm and compared with Page Rank algorithm and Twitter Rank algorithm, the algorithm obtained higher accuracy, lower complexity, and it can realize the dynamic update ranking.