Spectral Methods for Automatic Multiscale Data Clustering
CVPR '06 Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 1
A tutorial on spectral clustering
Statistics and Computing
Active learning for node classification in assortative and disassortative networks
Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining
Hi-index | 0.00 |
Overlapped communities detection in complex networks is one of the most intensively investigated problems in recent years. In order to accurately detect the overlapped communities in these networks, an algorithm using edge features, namely SAEC, is proposed. The algorithm transforms topology graph of nodes into line graph of edges and calculates the similarity matrix between nodes, then the edges are clustered using spectral analysis, thus we classify the edges into corresponding communities. According to the attached communities of edges, we cluster the nodes incident with the edges again to find the overlapped nodes among the communities. Experiments on randomly generated and real networks validate the algorithm.