Research On Mining Relevant Microbloging Users Of Specific Organization

This paper studies a new topic about mining the relevant users of the specific organization. It would dig out the user groups which has objective attribution or other affiliated with a particular organization from the virtual social networks. The research is very important to the mapping from the virtual network to reality. The main results of this study are as follows:First, this paper analyzes the generality and specificity of the new issue mining the specific organization associated user, emphatically analyzed relevant user mining method based on analysis of link relationships between users. This article attributed the special organization associated user data mining tasks to several sub-tasks as follows:the user crawling, digging the relevant user and analyze relevant user micro-blog. This paper focused on crawling the micro-blogging user data, micro-blog data and digging related user to the specific organization.Second, the paper designed and implemented a baseline system on mining associated users of specific organization based on link analysis techniques. The system was applied on sina-microblog platform and related experiment was carried out while regardding Beijing University of Posts and Telecommunications as the specific organization. Through the analysis of the system results, the article found that link analysis technology in this particular problem on the related user mining can be improved.Third, the paper presents a improved link analysis model based on word activation force method. Concept user active force was proposed base on the concept words activation force and fact as effective factor on the authority transmission of nodes in link analysis. Based on this algorithm, the organization related user mining contrast experiment was done, and the experimental results show that the algorithm improved the link analysis technology to the accuracy of relevant user mining, verified the improvement effect of WAF analysis technology to the link analysis techniques.Fourth, the paper designed and implemented a WEIYOU system of gathering, processing and analysis information for a specific organization for sina-microblog. First, the system overcomes the limitations, permissions of API and inconsistencies between API data, web data and other issues, access to information and user information from sina-microblog. And in information processing and extraction stages the paper solved the problem that label adhesion entity extraction. Then the system carried out data mining of Beijing University of Posts through the relevant mining and related microblogging users query expansion. Finally the system got both the relevant user and hot topics of Beijing University of Posts and Telecommunications.