Community-Based Link Prediction in Social Networks

Link prediction has attracted wide attention in the related fields of social networks which has been widely used in many domains, such as, identifying spurious interactions, extracting missing information, evaluating evolving mechanism of complex networks. But all of the previous works do not considering the influence of the neighbors and just applying in small networks. In this paper, a new similarity algorithm is proposed, which is motivated by the herd phenomenon taking place on network. Moreover, it is found that many links are assigned low scores while it has a longer path. Therefore, if such links the longer path has not been taken into account, which can improve the efficiency of time further, especially in large-scale networks. Extensive experiments were conducted on five real-world social networks, compared with the representative node similarity-based methods, our proposed model can provide more accurate predictions.