Link prediction in multiplex networks using a novel multiple-attribute decision-making approach

Abstract Recently, the problem of link prediction in multiplex networks has received increasing interest from researchers. Multiplex networks that model different types of relationships between the same group of nodes are a special case of complex networks. Studies have found that the structural features of different layers in a multiplex network are interrelated to some extent. Therefore, effective use of all layers’ information can improve the accuracy of link prediction. In this regard, we consider this problem as a multiple-attribute decision-making problem, in which alternatives are potential links in the target layer and attributes are diverse layers in the network. Moreover, we present a new multiple-attribute decision-making approach to solve the problem. To weight each layer in the proposed method, a layer similarity measure is defined based on cosine similarity. The performance of the proposed method is analyzed through extensive experiments. The results demonstrate that the proposed method can attain greater performance than competing methods in terms of accuracy and running time.