Discover Community Structure in Network by Optimization Algorithm Based on Modular Function

Abstract

Analyzing the structure of complex networks accurately and efficiently has become one hot topic due to the large-scale network in recent academic research. The existing optimization methods for community mining are mostly based on the function Q proposed by Newman. In this paper we introduce two complex network clustering algorithm models FN and spectral clustering. They are both aimed to maximize the value of function Q, the differences are that FN uses overall information and spectral clustering uses spectral graph theory. Then finally we apply these algorithms to analyze Chinese aviation network and come to conclude that Chinese aviation network is mainly composed of East-West and North-South routes, with which we can arrange the community structure.