Detecting Community in Complex Networks Using Cluster Analysis

Community structure is a common property that exists in complex networks. Improved spectral partition is used in this paper in order to transform the communities detecting into cluster analysis problem. Then, modularity function is applied to four cluster algorithms for community structure detecting. Especially, a new cluster genetic algorithm closely combined with modularity is proposed. We demonstrate the availability of our algorithms in three different kinds of network datum. We also make the comparison and analysis of the experimental results and obtain a conclusion that the proposed new algorithm presents fitness in initialization sensitivity and veracity. Finally, some further directions are pointed.