Community Network Partitioning Based On The Centrality Of Important Nodes

Nowadays,we are in an era of network.Complex network is ubiquitous in our life,and increasingly becomes the focus of people's attention and research.Through the study of complex networks,we can help people solve many practical problems,such as the prevention and control of infectious diseases,the spread of computer viruses,the elimination of power grid faults and so on.Community structure is one of the three characteristics of complex networks,and the other two are scale-free and small-world characteristics.Point connections within communities are relatively dense,and connections between communities are relatively sparse.Mining community structures is an important research direction of complex networks.Community structure can help us better understand the composition of the network,understand the operating mechanism of the network,and use the network to better serve human beings.Mining community structure is actually to divide the complex networks.At present,there are many community partitioning algorithms.In 2014,A.Rodriguez and A.Laio proposed a clustering algorithm based on density peak.The object of the algorithm is the points in multidimensional space.Under the premise of unknown prior information,the object is divided into a set of similar elements according to the similarity of objects.Inspired by this idea,we propose a community partitioning algorithm for complex networks based on the centrality of nodes.Therefore,the focus of this paper is to solve the following two problems:1.How to measure the density of nodes in complex networks to select cluster centers;2.How to measure the similarity between nodes,so that the nodes in the network can be allocated to the class where the cluster center is located.In view of the above two problems,the following work has been done in this paper:1.For the density problem of nodes,we take the edge of the graph into account and propose a weighted group density calculation formula,which can better measure the density of a node in the graph;2.For the problem of similarity between nodes,We use the most classical common neighbor similarity index(CN)to describe the similarity between nodes by considering the similarity measure of common neighbor index which is unique in the network;3.This paper presents a new community discovery algorithm in complex networks.By calculating the density of each vertex,the algorithm selects the cluster centers in the network according to certain rules,and then uses the CN similarity index to assign the remaining vertices to each cluster center.The algorithm does not need to know the number of clusters in advance,and avoids the problem that the number of communities k needs to be given in advance in most existing algorithms;4.This paper experiments on real network datasets and artificially generated network datasets to verify the effectiveness of our algorithm.