A citation similarity based community detection method in citation networks

Citation networks are important for us to understand the academic fields. By resolving the community structure, we can find out the subfields in the network. Many methods have been proposed to detect the communities in networks. However, they are not suitable to use directly in citation networks because they can be misled by some special papers and they do not take full advantage of the information contained in citation networks. To solve the problems, we propose a citation similarity based community detection method to detect the communities in citation networks. By transforming citation network to paper similarity network, we can use more information to resolve the community structure in citation networks and identify communities more precisely. The experiment results show that our method performs better in resolving community structure comparing with the method using directly in citation networks.