Link prediction in complex network based on modularity

Abstract

The problem of link prediction has attracted considerable recent attention from various domains such as sociology, anthropology, information science, and computer sciences. A link prediction algorithm is proposed using the modularity measure reflecting the community structure information of the network. Based on the fact that the connection likelihood between a pair of nodes in the same community is larger than those separated in different communities, we propose a new measure, named modularity contribution, for predicting link between a pair of nodes using information from intra-community and within-community of these nodes. Using the modularity contribution, we map the nodes to an Euclidean space. In this space, the nodes trending to be included in the same community are closely located. The cosine similarity of the nodes in this space is used in computing the similarity measure for link prediction. We also extend the method to solve the link prediction on networks of nodes with attributes. Our experimental results show that the proposed algorithm can obtain higher quality results than other algorithms.