Community Search Algorithm Based On Community Centrality

Many complex systems in real world,such as social networks,biological networks,power systems,can be modeled by complex networks.So the study on complex networks can help us gain a better understanding of real complex systems.One of the most important property exhibited by complex networks,and the one of most interest to this work,is community structure.Therefore,the community structure in complex networks has great significance.Traditional research on community structure is based on community detection,that is,to explore the whole network of community structure.However,the large scale and complicate structure makes it difficult to detect the whole community structure of the complex network.On the other hand,usually,researchers are more interested in the local community which are defined by the giving nodes.Therefore,the community search problem which focus on the local community becomes more and more important.Most of the existing community search algorithms define a goodness metric according to the statistics of community structure,with the goodness metric,these algorithms can get the target community by node clustering.This paper proposed a new community search algorithm CSIC which based on the community supporting degree.The main contributions of this paper is summarized as follows:1)Proposed the concept of search centrality.The vector representation of the nodes is obtained by the Skip-gram model,and the similarity between the nodes is calculated by using the vector similarity index.Finally,the reciprocal of the sum of the similarity of the node to the query node is taken as the query proximity of the node,and the query approximates the node the degree of relevance to the query node collection.2)A node weighting scheme is proposed,which combines the proximity of nodes and the importance of nodes.The attributes of node proximity and node importance are taken into account,and the reciprocal of weighted sum is used as the node Weight.3)On the basis of the node weighting model,the definition of the density in the social network analysis is given,and the density definition-community support degree with query bias is given.Community search algorithm CSIC based on community support degree is proposed,The CSIC algorithm avoids the problem of node clustering.Our experiments on real networks and synthetic networks demonstrate the accuracy and consistency of the new algorithm.