Research On The Key Technology Of The Influence Maximization And Propagation Prediction Of Microblog Message

With the rapid development of the Internet,the network has also developed into our daily necessities,and gradually change our behavior and lifestyle.Social networks have gradually replaced the traditional media,as people access to,the main platform for the propagation of message.Weibo as one of the most representative social networking platform,has become the mainfocus of social network research.In thispaper,based on the research of message influence maximization and message propagation prediction at homeand abroad,for the traditional message influence maximization algorithm and the message propagation prediction algorithm does not take into account micro-blogging network and micro-blogging users own characteristics and other issues,and strive to find a more effective influence maximization algorithm and More accurate prediction algorithm of message propagation.The impact of information maximization has been a hot issue in social network research.In order to maximize the influence of Weibo network news,it can effectively prevent large-scale proliferation of false messages in the field of public opinion monitoring,and can be more targeted in commercial applications of the advertising.In this paper,we consider the dynamic and high data volume of micro-blogging,and propose an influence-maximizing algorithm based on user activity and topic.The algorithm mainly considers the user 's topic and the user' s activity degree,and improves the forwarding probability of the independent cascade propagation model according to the user 's activity.Finally,the simulation and experiment are compared with the KDD CUP 2012 data set of Tencent weibo,which proves that the improved algorithm is better than several heuristic algorithms in message influence maximizing.A t the same time,the time algorithm's complexity is better than the greedy algorithm.The user active degree of this factor,but also can effectively shield the robot users and zombie users to maximize the impact of the adverse effects.Through the research of microblogging message prediction,it is predicted that the future trend of news can guide the dissemination of news better.The current research can be divided into two types: information-centric and user-centered.Information-centered predictive research often neglects the individual's communication behavior,focusing only on the overall communication trend.The main task of user-centered research is User communication behavior prediction,to provide personalized recommendations possible,according to the user attribute specific analysis of whether the user will participate in the dissemination of a message.In this paper,we study the overall trend of message propagation.Aiming at the current algorithm ignoring the user-related attributes,an algorithm based on user quality is proposed for message propagation prediction.The results of experiments show that the proposed algorithm outperforms other algorithms.