Abstract:Link prediction plays a significant role in various applications of complex networks. The existing link prediction methods can be divided into two categories: structural similarity algorithms in network domain and network embedding algorithms in the field of machine learning. However, few researchers focus on comparing these two categories of algorithms and exploring the intrinsic relationship between them. In this study, we systematically compare the two categories of algorithms and study the shortcomings of network embedding algorithms. The results indicate that network embedding algorithms have poor performance in short-path networks. Then, we explain the reasons for this phenomenon by computing the Euclidean distance distribution of node pairs after a given network has been embedded into a vector space. In the vector space of a short-path network, the distance distribution of existent and nonexistent links are often less distinguishable, which can sharply reduce the algorithmic performance. In contrast, structural similarity algorithms, which are not restricted by the distance function, can represent node similarity accurately in short-path networks. To address the above pitfall of network embedding, we propose a novel method for link prediction aiming to supplement network embedding algorithms with local structural information. The experimental results suggest that our proposed algorithm has significant performance improvement in many empirical networks, especially in short-path networks. AUC and Precision can be improved by 36.7%-94.4% and 53.2%-207.2%, respectively. Published under license by AIP Publishing.

A Survey of Link Prediction Algorithms

Network embedding for link prediction: The pitfall and improvement.

Link Prediction in Complex Networks: A Survey

A New Link Prediction Algorithm Based On Local Links

A Survey on Hyperlink Prediction

Experimental Analyses on 2-Hop-based and 3-Hop-based Link Prediction Algorithms.

Link Prediction Based on Local Structure and Node Information along Local Paths.

Link Prediction on Complex Networks: An Experimental Survey

A Survey of Link Prediction in Information Networks

Empirical Analysis of Traditional Link Prediction Methods

A Link Prediction Method for Large-Scale Networks

A similarity index algorithm for link prediction

Link Prediction Via Linear Optimization

Synthetic graphs for link prediction benchmarking

A New Link Prediction Algorithm: Node Link Strength Algorithm

Link Prediction Based on Node Similarity

A New Prediction Algorithm Based on Local Neighbor Closeness and Influence of Node.

A Link Prediction Algorithm Based on Weighted Local and Global Closeness

Predicting Missing Links Via Correlation Between Nodes

Link Prediction Algorithm Based on the Initial Information Contribution of Nodes

Local Structure Similarity Algorithm for Link Prediction