Link Prediction in Complex Networks Based on a Hidden Variables Model

The study of complex networks has attracted the attention of many research in recent years due to their potential to describe complex systems. Identifying potential missing links in networks has a wide range of applications. In this paper, we propose a new approach for solving link prediction problem based on a hidden variables model. Recent studies suggest that real networks have underlying hidden metric space that explain their structural characteristics. We suggest to model the link prediction as an inverse problem in order to infer node coordinates in a hidden metric space and subsequently use the produced model to find the missing links and detect the spurious ones. Our results show that the proposed approach can provide more accurate prediction than other existing methods.