Research On Community In Singed Network

With the continuous progress of the era of science and technology,the complex systems in real life are abstract into complex networks,and the complex analysis of complex network structure has attracted great attention in recent years.As an important part of the complex network,the study of the community structure is of great significance for the network analysis.Finding the community structure in complex networks provides an important basis for solving classical network problems such as data mining,collapse problems and site selection problems.Studying and understanding the community structure of complex networks is helpful to solve many practical problems.In recent years,researchers have proposed many community detecting algorithms,but most of them only consider the density of edges in the network,and only apply to unsigned networks.However,in real life,there are a lot of social relations expressed by signed network.In the community division of signed network,we need to consider not only consider the density of the network edges but also the symbols of the network edges.Therefore,we propose a community division algorithm for the signed network.The community detecting algorithm in this paper is mainly divided into two stages:in the first stage,according to the structural properties of the signed network,we redefine the density and distance in a signed network to characterize the clustering center where the nodes are redifined similarity.The proposed algorithm generates a decision graph based on the density and distance of the nodes of the network,the clustering centers are determined easily.In the second stage,our goal is to minimize the structural conflict existing in the community division by adjusting nodes from one community to another.Iterations will repeat until the number of structural conflicts be stable.We conducts the experiment on the artificial generated signed network and the real signed network,and compares with other community division algorithms.Based on the classical community division evaluation index,we verify the accuracy and effectiveness of the proposed algorithm.