Research On Similar Users Discovery Of The Online Microblog Network

With the popularity of the Internet,the mainstream of online social media has made rapid development,and micro-blog in these online social media has shown a more prominent development trend.Micro-blog has attracted a large number of people.These people share their views,hobbies,ideas and other information on micro-blog.Therefore,micro-blog has huge impact on the development of society.In recent years,researchers at home and abroad have done a full range of research based on online social media data,such as community discovery,commodity recommendation,similar user recommendation,etc.This thesis put its focus on the micro-blog,carries out the researches of theory and methods on the topic of individual mining and similar user mining.Specific research contents include the following three aspects:(1)According to the characteristics of dynamic micro-blog social,this thesis use timeslice to divide the information of user's micro-blog,then mining dynamic topic of users.This method can better reflect the dynamics of micro-blog users' topic,and get more accurate mining results of similar users.(2)Based on the comprehensive analysis of micro-blog network user's characteristics,this thesis purposes a new method for user 's similarity calculation from three aspects: the similarity of user background information,the similarity of user 's dynamic topic,and,the similarity of user 's social relations.After calculating three similarities,this thesis obtains the final similarity between two users through the weighted method,and then finds the similarity between the specific users and users set climbed out from micro-blog.By generating the descending rank queue based on the size of these similarities,top-k users will be recommended to specific users.(3)In order to display the research results intuitively,this thesis designs and implements the prototype system of similar users mining based on the micro-blog according to the researches of the first two parts.This system has many functions,such as dynamic topic mining,similar user mining and so on.The researches of this thesis can help analyzers to find similar users of a specific user of microblog,expand their interpersonal relationship,and find more interesting topics for them.Also,it will be useful for the users to find their role in the social network,and do the data support for the community discovery.