Edge-content based community detection algorithm on email network

As email playing an increasingly important role in the online world, analyzing and mining on email communication network draw more and more attention, in which community detection is one of the most key technologies and applications. Based on the intuition that the edge topic is more focused than the node topic in a short period, this paper proposes an edge-content based email network community detection using clustering algorithm. First extract the related emails on the edges, and feature terms are extracted from those emails in combination with mail body and subject to label the edge-content. Express the edges of a network with Vector Space Model (VSM), and then cluster the edges according to their contents by an advanced k-means algorithm to obtain community. Afterwards the obtained communities are described and evaluated. Experiments show that the proposed method is promising.