An Eigenvector-Based Kernel Clustering Approach to Detecting Communities in Complex Networks
To detect communities in complex networks, we generalize the modularity density(D) to weighted variants and show how optimizing the weighted function(WD) can be formulated as a spectral clustering problem, as well as a weighted kernel k-means clustering problem.We also prove equivalence of the both clustering approaches based on WD in mathematics. Using the equivalence, we propose a new eigenvector-based kernel clustering algorithms to detecting communities in complex networks, called two-layer approach. Experimental results indicate that it have better performance comparing with either direct kernel k-means algorithm or direct spectral clustering algorithm in term of quality.
KeywordsSpectral Cluster Community Detection Kernel Matrix Normalize Mutual Information Community Versus
Unable to display preview. Download preview PDF.