Research On Community Detection Algorithm Based On Second-order Neighbor Importance

Association detection is a hot research topic in the field of complex network analysis,and its goal is to find the set of nodes with strong internal connections but sparse external connections,so as to explore the topology of associations and the hidden relationships between functional units.Seed extension-based association detection algorithms are widely used to solve this problem because of their high efficiency and effectiveness,however,some of them need to rely on a large amount of a priori information in real networks leading to their failure to achieve good results in terms of accuracy and quality.To address the above problems,this paper combines node influence and node similarity to classify associations,and the main contents are as follows:(1)In response to the current seed expansion algorithm,which usually only considers the attributes of the nodes themselves or randomly selects nodes as seeds when selecting core nodes,which may lead to unstable performance of the algorithm,this paper proposes a seed expansion algorithm based on the second-order neighbor importance.In this paper,we first study the factors that influence the node’s second-order neighbors and combine their own attributes with their first-order and second-order neighbors and node information entropy to calculate node influence to efficiently obtain the set of core nodes in the network;association expansion is carried out in two steps,the first step divides the neighbor nodes that are connected to only one seed node,the second step uses the node belongingness index to divide the unassigned nodes in the network to further improve the accuracy of the algorithm;finally,this paper optimizes the associations using node belongingness and modularity metrics,and this process can merge small associations that are not realistic,so as to obtain a more accurate association structure.In addition,the algorithm is highly adaptable and can be applied to detect associations in different types of networks and does not require any a priori knowledge of network information.(2)Since overlapping associations are more in line with real-world complex networks than non-overlapping associations,in order to perform the detection of overlapping associations,this paper extends the previous algorithm and proposes an overlapping association detection algorithm based on second-order neighbor importance,which inherits the seed selection strategy from the previous algorithm and selects the set of high-quality seed nodes that can represent associations in the network;the association extension uses an improved In this process,a node can belong to multiple associations at the same time,and if there are nodes that do not belong to any association,a node belongingness metric is introduced to divide the nodes twice.Finally,the association distance metric is introduced to find overlapping associations and merge overlapping associations to eliminate excessive overlap and redundancy,thus optimizing the association division structure.Experimental verification and comparative analysis are carried out on multiple sets of artificial and real network datasets.The results show that the proposed two community detection algorithms outperform the comparison algorithms in terms of performance,and achieve better results in terms of accuracy and quality.