The Research On Community Clustering Algorithm Based On Node Density

Networks exist in all aspects of our lives,such as the Internet,transportation networks,telephone networks and other technical networks,as well as biological networks such as neural networks and ecological networks,as well as social networks and information networks.Research networks can deepen our understanding of the real world.As for the network,the community structure is a very important feature,which indicates that the connections among intracommunity nodes are dense,while intercommunity nodes are sparse.The process of finding out the community structure in the network is called community detection.Community detection helps us understand the topology of the network and mine potential information in the network.Therefore,the research community detection algorithm is very significant.A large number of community detection algorithms have been proposed in complex networks.The k-means clustering algorithm is widely used in community detection due to its advantages of fast clustering and easy implementation.This paper studies the relationship between nodes and community structure in the network,proposes the concept of node density,and applies it to k-means clustering algorithm,and proposes a community clustering algorithm based on node density.This algorithm avoids the problem of random selection of the initial clustering center in the traditional k-means clustering algorithm.The appropriate correlation matrix is defined by the node similarity,which reduces the number of iterations of the k-means clustering algorithm.The experimental results prove the proposed algorithm can find real communities more accurately.The main work of this article is as follows:1.This paper studies the certainty relationship between nodes and community belongingness in the network,Community Belongingness Uncertainty(CBU)of nodes is defines by introducing information entropy,and proposes the concept of node density,which is used to describe the community belongingness of nodes.2.We solve the problem of random selection of initial clustering centers in traditional k-means clustering algorithm.In this paper,node density and node degree are combined to balance the centrality and community attributes of the node,and it is used as the index DD(the combination of node density and node degree centrality) for the initial node selection of the k-means clustering algorithm.3.Using node similarity to define the correlation matrix between nodes,which is used as the input of k-means clustering algorithm to iterate,and output the result community.The algorithm proposed in this paper is compared with the traditional k-means clustering algorithm and some existing algorithms on real networks and synthetic networks.The experimental results show that the proposed algorithm is feasible and has good accuracy,especially for the community detection of big data sets,and has positive reference significance for the application scenarios based on community detection.