Communities are of great importance for understanding graph structures insocial networks. Some existing community detection algorithms use a singleprototype to represent each group. In real applications, this may notadequately model the different types of communities and hence limits theclustering performance on social networks. To address this problem, aSimilarity-based Multi-Prototype (SMP) community detection approach is proposedin this paper. In SMP, vertices in each community carry various weights todescribe their degree of representativeness. This mechanism enables eachcommunity to be represented by more than one node. The centrality of nodes isused to calculate prototype weights, while similarity is utilized to guide usto partitioning the graph. Experimental results on computer generated andreal-world networks clearly show that SMP performs well for detectingcommunities. Moreover, the method could provide richer information for theinner structure of the detected communities with the help of prototype weightscompared with the existing community detection models.
展开▼
