Community Detection of Chinese Micro-Blogging Using Multi-Dimensional Weighted Network

Existing community detection methods are mostly based on the analysis of the links among the nodes, ignoring the rich, while the others often ignore the network structure which is the foundation of social media. Aiming at the existed problems, this paper proposed a community detection algorithm based on multi-dimensional weighted network. By introducing User Interactive Frequency, User Interest Similarity, and User Attributes Similarity into the basic network topology, a multi-dimensional weighted network is set up. After converting the multi-dimensional weighted network into a single-dimensional weighted network, an improved CNM algorithm is exploited to discover the communities. A corresponding series of evaluation indicators are proposed to evaluate the detection results. By evaluating the algorithm in the dataset of Chinese Micro-blogging, it reveals that the clustering results are better when extra information is used, and in Chinese Micro-blogging platform, User Interactive Frequency plays a much more important role in community detection.