Research On Community Enhancement Detection Algorithm Based On Graph Compressio

In recent years,with the rapid development and transformation of information and communication technology,the real world is full of a large number of rich and valuable complex networks,and research on complex networks has also attracted more and more attention from researchers.Community structure is an important feature in complex network,and it has important value and significance for in-depth study of complex network structure and functional characteristics.Through community discovery of complex networks,the hidden group structure in the network can be explored,and a certain structure or hidden associated information contained in the real data set of the network can be revealed,which has important practical value and has been widely used in personalized recommendation.,public opinion analysis,advertising,information retrieval and other fields.Community detection is one of the important tools for discovering useful information hidden in complex networks,which is often difficult to obtain through simple observations.With the development of The Times,the accumulation of data cause our current network,more and more nodes,will make the number of edge in the complex network grow exponentially,which will lead to unclear network community structure,make most of our existing community detection algorithm on complex network detection accuracy is not high,and in a larger scale of complex network,our traditional detection algorithm time cost is higher or even cannot get the results.In order to deal with this problem,this paper proposes a community detection algorithm(CSEP),which can improve the accuracy of community detection while reducing the network size.The algorithm mainly includes two aspects:The first step is graph compression.This step can greatly improve the scalability of the community detection algorithm.This paper presents a simple,efficient,and lossless method for social network graph compression.The proposed method can greatly reduce the storage space and computational time for community detection.Specifically,we iteratively merging vertices with a degree of 1 or 2 into their neighbors with a higher degree.The neighbor nodes we define as a supernode.For the nodes and hypernodes with degree 1 or 2,we delete the remaining edges and store the weights of the edges in the hypernodes through the formula,recalculating the edge weights of the hypernode and its neighbors,so that we reduce the size of the network.After this step,we obtain the compressed complex network.The second step is the community structure enhancement.We enhance the community structure of the network by adding the links between nodes that may belong to the same community to reduce the links between different community nodes,thus transforming the fuzzy community structure into a clearer structure than the original.Experimental results show that the proposed community detection algorithm outperforms five state-of-the-art community detection algorithms on real networks.