Node Content And Link Relationship Between The Microblogging User Interest Community Found

With the rapid development of Internet technology, the model of social networks has transformed from the previous real network into a virtual network on the Internet. Microblogging as a new social network, the characteristic of short content, published fast and a variety of forms just meet the need of real-time information obtaining and convenient for people. It’s has great significance of finding social network community structure, friends recommending and microblogging marketing through analyzing of social networks.Traditional community method generally equal the microblogging social network as the traditional social networks, ignoring the attribute information of individuals. In microblogging network, the user generally browsing and accessing information based on interests, the user’s microblogging information, attention information, and even the basic information are all reflecting the users’ interest. As the saying goes, "people with class together, things in groups", if we can find user groups in a common interest, then it will provides a new entry point for friends’ recommendation and microblogging marketing. This paper will do research on microblogging users interest community discovery based on the characteristics of microblogging network and combined with traditional methods of community discovery, the main research work includes the following aspects: 1) The research of microblogging data acquisitionUse the methods of Web Browser control to realize microblogging data acquisition. By simulating a user logs on, respectively from the user’s microblogging information pages, following pages and fan pages crawl the microblogging data, and then use regular expression matching data to obtain the required data. 2) The research of mining opinion leadersPropose a method of mining opinion leaders by using the microblogging property information and relationship information. By analyzing the characteristics of microblogging users, respectively construct the initial impact compute model based on the user attribute information and impact diffusion model based on the users’ relationship. And gives quantitative calculation method of users’ impact. By threshold analysis to determine the final opinion leaders. 3) The research of interest community discovery based on users’ contentPropose a method of interest community discovery based on K-means clustering algorithm. By analyzing the different characteristics of opinion leaders and ordinary microblogging users, respectively construct the opinion leaders interested models and general users interest models. On the basis of the acquired interest information and use the K-means clustering algorithm realize interest community discovery. 4) The research of community discovery based on link relationshipPropose a local community discovery algorithm based on the coefficient. By analyzing the structural characteristics of complex networks, defines the concept of point aggregation, divide groups of nodes by the concept of clustering coefficient and edge coefficient, and ultimately realize the community discovery. 5) Design and implementation the system of microblogging users’ interest community discoveryUsing the windows forms application implements the design and implementation the system of microblogging users’ interest community discovery. It describes the system overall framework and the respective sub-modules function.On the basis of platform constructed system, validate the effectiveness of the above mentioned methods by experiment.