Knowledge Discovery Research Of Chinese Microblog User Behavior

With the development of WEB 2.0,popularization of mobile internet,and the promotion of the public information literacy,microblog has become an important platform for creating,propagating and obtaining information due to the advantages of convenience,celerity and accessibility.Nowdays microblog is widely used in our daily life,work,learning and social interaction.As we have millions of microblog user behavior data,we can analysis the data and discover the knowledge by data mining technique to analysis user need,search for authority user,and discover microblog theme,which will contribute to microblog resource utilization,knowledge service,crisis communication and promote social harmony and stability.Knowledge discover based on Microblog user Behavior is a comprehensive research topic:the data source and format of user behavior data are many and varied,new knowledge discover model and algorithm has been proposed continually,user habit and grammatical structures are different at home and abroad,and Chinese microblog is still undergoing rapidly-evolving progress.Based on the foundation theory and review of related literature,my research classifies the microblog user knowledge model into three categories.the first kind of models make use of the statistical data.The statistical significance can reflect the characters of microblog user and context,discover the regular pattern of social network behavior,classify user and predict the flow of information through networks.The second category of knowledge discover model focuses on the analysis of user relations within microblog.As a social network media,tacit knowledge,such as user interest,preference and intentions.By the methods of social network analysis,this article propose models to discover the community structure,finding influence user and personalized recommendation.The third category of knowledge discover model is based on the connection of the microblog user and context.My research analysis the context which user reposts,comments and mentions by means of word segmentation and text mining to find users cluster,mocroblog theme and keywords.This thesis takes the mention behavior,for example,according to the mentioned microblog user and relative context,user clustering and label extractingToword the above three knowledge discover models,this article collected three sina weibo user behavior datasets,propose corresponding models.In the last part,this thesis construct a general model for knowledge discover based on user behavior,which include the process of data prepration,pateran and knowledge repository management.This thesis also introduce several knowledge service methods,which can be used for guiding the research and application of the new knowledge demand.