A Link Prediction Algorithm Based on the Closeness of Nodes in Complex Network

Abstract

Link prediction, as one of the social networking research questions, predicts the possibility of
links between nodes that are not currently linked by existing information such as node
information and the characteristics of the entire network topology. In terms of quantity, adjacent
nodes in the topological information considered by the local information algorithm do not take
into account the interrelationship between them and the close relationship between the small
groups. In this paper, an algorithm based on compactness similarity is proposed to settle the
dispute. By improving the calculation steps of the similarity of adjacent nodes, it preserves the
simplicity and efficiency of the traditional local similarity calculation steps of local information,
and makes make the best of the structure of small groups of the network to guide the prediction,
so that the actual performance of the link prediction is more effective.