Social Networks Influence Analysis Based On The Social Features

Along with the high-speed development of Internet technology, online social network has become the important medium of communication contemporary. Since the advent of the we media age, network provides great demend for measuring user influence, and created favorable conditions for the study which use Internet technology to the accumulation of user data.The study of the influence of the social network has important application value in the information dissemination, public opinion control, product marketing,and is one of the hot spot of current research.However,we found some research ignore the social character of the users on influence in the process of summary and analysis of related work.Therefore, innovation in combining the theory of sociology, this dissertation analyses the social network user information dissemination rule, is presented based on the users SRank social characteristics influence evaluation model. Specific work is as follows:(1) Through analyzing characteristics of online social networks, introducing online social network information transmission for important elements as well as the related patterns of information transmission, analyzes the form of user's influence and generate the required environment foundation.(2) Analysis the deficiencies of the influence of traditional model in applied on current online social network, PageRank, for example, in view of the three aspect which was ingnored by the traditional influence models, and summarize the rule of online social network information transmission, through the network user data.(3) Based on the social network information transmission law, combined with relevant theory of sociology, the SRank user influence model was summarized, and based on this, evaluation algorithm which measure the user influence by predicting capability of user information dissemination was proposed.Under the same data set, compared with PageRank and its improved algorithm, the algorithm obtained better prediction of influence, SRank user influence evaluation of the effectiveness of the algorithm was verified.