Research On Hidden Key User Discovery Technology In Microblog

Microblog is a double-edged sword,which not only facilitates people's information interaction,but also the spread of malicious public opinion.Many malicious public opinions' initiating users in Microblog are clandestine,which increase the difficulty of the supervision of public opinion in Microblog.Therefore,identifying the hidden key users is of great significance to the media supervision and guidance in Microblog.This article takes Microblog users as the research object,and conducts in-depth research on hidden key users in Microblog.The main work of the full text is described as follows:1.Based on the concerning relationship in Microblog,the relationship network is constructed,and various interactions carried by the network are analyzed in depth.The definition of hidden key users and significant ones in Microblog network is given,and a research framework of finding hidden key in Microblog network is proposed.The user's research framework establishes a foundation for the following research.2.For the problem that the user's blogs that don't possess the disseminating quality will interfere with the user influence analysis,thus the characteristics of the disseminative blogs are studied in depth,based on which we filter the blogs,and then,combined with the fans' attention to the user and the topic similarity between both of them,calculates the user's blogs' spreading range expectations and uses this as an indicator to analyze the user's influence.Experiments show that the method has a high accuracy.3.Key users can be distinguished as significant ones and hidden ones.Existing methods mostly analyze the direct influence of users,thus most key users discovered by them belong to significant key users.However,although the hidden key users have high influence,yet their concealment is also very high.They generally influence the significant key users to create impacts.For this phenomenon,based on the transfer entropy,the hidden influence of normal users on significant key users is measured,then combine the direct influence of the two to comprehensively assess the hidden importance of normal users.The experiment shows that this method has good accuracy.