A Novel Similarity-Based Method for Link Prediction in Complex Networks

In complex systems with interactive elements, link prediction plays an important role. It forecasts future or missing associations among entities of a complex system using the current network information. Predicting future or missing links has a wide variety of application areas in several domains like social, criminal, biological, and academic networks. This paper presents a novel method for finding missing or future links that uses the concepts of proximity between the vertices of a network and the number of associations of the common neighbors. We test the performance of our method on four real networks of varying sizes. We tested it against six state-of-the-art similarity-based algorithmss. The outcomes of the experimental evaluation demonstrate that the proposed strategy outperforms others. It remarkably improves the prediction accuracy in considerable computing time.