The User Analysition Based On The Social Media

Hundreds of millions of users exist in the social media represented by Sinamicro-blog, through which they publish, communicate and review contents. Whilegenerating a lot of contents, the relationship between them is also expanding. Theentire social media is connected by user entities whose offline behaviors aregradually migrated to the online. Therefore, how to analyze user entities throughtheir online behaviors seems particularly significant.Figures in the Micro-blog media will generate relationships and contents intheir process of reaction. Relationship analysis can dig out the user’s close friends,alumni and colleagues. Content analysis can be used for topic identification, topicclassification and interest mining, etc. As for relationship analysis, this paperintroduces a community partition algorithm on the basis of organization identity.And, for topic identification, this paper employs a method of similarity degreecalculating on the basis of topic model.Analysis techniques require applications of scenes. Techniques of characteranalysis are mainly used to recommend characters and information. This paperanalyses the characteristics of the data of characters in the Micro-blog media,proposing a multi-dimensional character recommendation algorithm which isproved more accurate than the traditional recommendation algorithm throughexperiments. And this paper will elaborate information recommendation algorithmfrom three aspects: length of micro-blog, relativity between micro-blog and users,familiarity between users. Eventually, based on character analysis, this paperdevelops an application system “character analysis system of Sina micro-blog”.