Weighted network overlapping community partition based on node membership

Abstract

Aiming at the problem of random selection of core nodes in the existing algorithm of overlapping communities, which leads to low accuracy of community division, this paper is based on the idea of optimizing node selection and then accurately dividing overlapping communities, this paper proposes an algorithm of dividing overlapping communities based on the degree of membership for the weighted network. Firstly, the common neighbor nodes are introduced, and based on the overlap ratio of node neighborhoods, the node overlap strength and node unit overlap strength are defined. Then select the core community, it can objectively reflect the importance of the core community in the networks9; Secondly, the degree of membership of nodes relative to the core community is calculated, and the network overlapping community is preliminarily divided. Finally, the extended module degree is used to optimize the sub-communities initially divided to realize the overlapping community division. Experiments show that the algorithm improves the accuracy of overlapping community partition.