Abstract:The link prediction algorithm which based on node similarity is the research hotspot in recent years. In addition, there are some methods which based on the network community structure information to predict the missing links, however, these studies only concerned about the obvious information between different communities such as direct links. We found that it is hard to predict the missing links if the two communities have little direct connections. In fact, there is similarity between communities such as the similarity between nodes and this similarity is significant for prediction. So, we define a community similarity feature which named community relevance by using not only the obvious information but also the latent information between different communities in this paper. Then a novel algorithm which based on the community relevance and ruler inference is proposed to predict missing links. In this method, we extract the community structure by using the local information of the network first. Next, calculate the relevance of each pair of communities by using the new community relevance indices. Finally, a simple prediction model which based on ruler inference is applied to estimate the probability of the missing links. It is shown that the proposed method has more effective prediction accuracy and the community relevance features improve the predictor with low time complexity, with experiments on benchmark networks and real-world networks in different scales, and compared with other ten sate of the art approaches.

Cliques in Complex Networks Reveal Link Formation and Community Evolution

The Reconstruction of Complex Networks with Community Structure

Community Detection in Complex Networks using Link Prediction

Link Prediction in Complex Networks: A Clustering Perspective

Effective Model Integration Algorithm for Improving Link and Sign Prediction in Complex Networks

Community Detection in Complex Networks with an Ambiguous Structure Using Central Node Based Link Prediction.

Link Prediction in Complex Networks: a Local Naïve Bayes Model

Link Prediction in Complex Networks: A Mutual Information Perspective

Community and Local Information Preserved Link Prediction in Complex Networks.

Prediction of Missing Links and Reconstruction of Complex Networks

Accurate and Fast Link Prediction in Complex Networks

Prediction of Missing Links Based on Multi-Resolution Community Division

Prediction of Missing Links Based on Community Relevance and Ruler Inference

Predicting missing links and identifying spurious links via likelihood analysis

An Information-Theoretic Model for Link Prediction in Complex Networks

Link Prediction Via a Neighborhood-Based Nonnegative Matrix Factorization Model

Efficient link prediction model for real-world complex networks using matrix-forest metric with local similarity features

Prediction of link evolution using community detection in social network

Link Prediction Based on the Mutual Information with High-Order Clustering Structure of Nodes in Complex Networks

Toward Link Predictability Of Complex Networks

Multi-Resolution Prediction Model Based on Community Relevance for Missing Links Prediction.