A Micro-blogging Community Discovery Method Based On User's Interactions

As a new social networking states, micro-blogging gets user's favor with its unique advantages, has become one of the mainstream social networking states in a few years. In the micro-blogging network, users with the same or similar interests belongs to the same community. Community discovery aimed at dig out the hidden community structure, divide micro-blogging users to community, the study has important theoretical value and application value: in theory, this study can deepen our understanding of micro-blogging network structure and rules; in application, the study can be used for goods recommendation, advertising push, friends recommend and other services.In micro-blogging network, there are two types of user's contraction: Based on the follow mechanism and based on the interactions between users. According to this, the community structure of micro-blogging network is divided into "follow relationship micro-blogging community" and "interaction relationship micro-blogging community". Through in-depth analysis of micro-blogging network structure and the interactions between the user's contribution of community, found that the "interaction relationship micro-blogging community" more able to reflect the real micro-blogging network community structure. In order to find the community structure, first, building user's interaction behavior modeling, according to the true that user posted posts and interactive posts all can reflect user's interest, designing algorithm to calculate specific values of the weight of the interaction behavior. Designing the calculation of micro-blogging user's interest to posts, then, dividing the interactive relation between micro-blogging users to direct interaction and indirect interaction referencing author cross citation analysis and author coupling analysis in citation analysis, proposing their calculation formula based on user's interest to posts calculation formula. And then put forward micro-blogging user's interaction similarity calculation and micro-blogging network community discovery based on users interaction similarity. Final, Experiments on Sina Weibo use function to evaluative the results, verify the validity and rationality of the algorithm.