Local community detection algorithm based on links and content

Community detection is an important field in research of social networks. There exist a lot of algorithms which most of them are based on the density of connections between groups of nodes. On the one hand, the error and lack of links may lead to great impact on the result of community detection. On the other hand, there are users with deep relation but without much communication, so the density of connections can't represent whether the users belong to the same community or not. With the network becoming more and more complicated, the traditional global method will cost much time and space. In this paper, we proposed a local method based on links and content, and the method focuses on particular users' communities. The results on Enron email dataset have shown the superior performance and accuracy rate of our proposed method in community detection.