Community structure is an important feature of complex networks, and it is of great significance for us to understand and analyze other characteristics of the network, meanwhile it also helps to identify the properties of individual nodes by extracting community structure within the network. In this paper, a novel community detection algorithm based on the node correlation strength is proposed, where the edge and weights of the node is used to calculate the correlation strength of each node in the network, and then search the higher modularity by moving a node with a low correlation strength to a neighbor's community. After that, we fold nodes within the same community to reconstruct the network for a new node and then recursively implement this process to obtain the best partition scheme with the higher modularity. Finally, the algorithm is applied into computer generated network and real networks and compared with the existing algorithms. The current results show that the partition scheme given by this new algorithm has the higher modularity, it also indicates that the number of communities is consistent with ones within several realistic networks.
展开▼
