Community Detection Research Based On Signed And Dynamic Network

Complex network system exists in our lives,affecting our daily life.The vast majority of networks have clustering characteristics with community structure.How to find the community structure of real network system has become a social concerns quickly.To solve the problem of community detection,the detections of community in signed network and dynamic network was considered in this paper:(1)Community detection in signed network: a dynamic evolutionary community detection algorithm based on intimacy in signed network has been proposed in this paper.A new similarity function is added,and the similarity function based on shortest paths have been added in order to give the similarities of nodes without link.A dynamic model is constructed for dynamic evolution of intimacy in signed networks.Finally intimacy of the nodes in the same communities are equal to 1,and the intimacy of nodes in different communities are equal to 0 over time.Networks will be divided into several clusters based on the improved model.In order to verify the performance of the proposed method,the algorithm has been tested by USC network,GGS network and 17 Synthetic networks.The results showed that our proposed algorithm is efficient.(2)Community detection in dynamic network: based on the algorithm of community detection in signed network,time factor is also considered.Firstly the similarity between previous time and next time of the network is weighted.Then the network model is improved,so that nodes with high intimacy are in same community.Finally the network is divided into different communities and validity of the algorithm is tested by real network and synthetic network.In order to verify the effect of weighting factorα on the algorithm,different values of α are compared and find that parameter α has no effect on the algorithm.