Research On Algorithms Of Community Detection In Complex Networks

Complex systems in the real world can be highly abstracted from complex networks,and the community structure is the most significant characteristic of complex networks.By mining the community structure of complex networks with the algorithms of community detection,the relationship between the network topology and the hidden characteristics can be deeply explored.Besides,research results of community detection in the network have been widely applied in many fields,such as criminal gang identification,protein complex analysis,personalized recommendation and information retrieval.In recent years,a large number of community detection algorithms have been proposed in the literature to divide the community structure in the network.However,most of these algorithms cannot achieve good results in terms of accuracy and efficiency.Based on the analysis and elaboration of the existing community detection algorithms,this paper proposes two novel community detection algorithms for the existing problems.The main research work and contributions are as follows:(1)In order to improve the accuracy and stability of the label propagation algorithm LPA,this paper proposes a novel node gravitation-based label propagation algorithm.The algorithm first calculates the importance of all nodes in the network based on the LeaderRank algorithm of the global characteristic of the network,and determines the update sequence of the nodes in ascending order of importance.Then,the algorithm combines the topological characteristics of the network and the gravitational theory to define the virtual gravitation between nodes in the network.The definition of gravity also takes into account the node characteristics of node importance and node similarity.In the process of label iterative updating,the label with maximal total gravitation are selected to update the label of the current node in order to improve the stability and accuracy of the algorithm.(2)Aiming at the problem of community detection in large-scale network,this paper proposes a community detection algorithm based on the belonging intensity of intermediate nodes,and divides the community structure of the network by analyzing the belonging intensity of intermediate nodes in different communities.Firstly,the algorithm defines the intermediate nodes of the network and their belonging intensity in the community according to the local topological features of the network.The definition of membership strength is based only on the local topological characteristics of the nodes in the network to improve the computational efficiency.Considering the prior knowledge of the network and the influence of different characteristics of nodes on their associations,adjusting parameters are added to the definition of belonging intensity to control the influence of different characteristics.Iteratively analyzes the belonging intensity of intermediate nodes in different communities to divide the community structure of the network.In order to improve the efficiency and accuracy of the algorithm,rough initialization method and post-processing method are also proposed.In order to verify the effectiveness of the algorithm proposed in this paper,the experimental of this paper uses some network datasets with different characteristics to compare the performance of the proposed algorithm and the comparison algorithm on multiple indicators.The experimental results show that the community detection algorithm proposed in this paper can effectively detect high-quality community structure.