Abstract:Evolutionary models are widely used to analyze the underlying mechanism of network evolution.However,judging the performances of different models is not easy because there are too many statistical features that should be taken into consideration.Actually,the aim of link prediction is estimating the likelihood of the existence of links,which is also what an evolving model wants to show.Therefore,link prediction is expected to provide a fair platform to better quantitatively compare different models.We firstly introduce the framework of proximity-based link prediction,and then discuss how to use link prediction to compare evolutionary models.Finally,we take the Chinese city airline network as an example to show the effectiveness of our method.When we use single factors(structural factor,geographical factor and some external factors like population and economic level) to predict missing links,the common-neighbor-based index gives the highest accuracy,which suggests that the evolution of the Chinese city airline network is mainly affected by structural factors.Furthermore,the hybrid method of common neighbors and the tertiary industry provides a much better prediction than any single algorithm or other hybrid methods,in accordance with our previous works on partial correlation and causal analysis.Our work provides a new perspective and analysis tool for the studying of network evolution.

A Community-Structure Based Adaptively Optimized Link Prediction Algorithm

Community Detection in Complex Networks using Link Prediction

Link prediction based on depth structure in social networks

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

Revealing how network structure affects accuracy of link prediction

Enhance Ambiguous Community Structure via Multi-strategy Community Related Link Prediction Method with Evolutionary Process

Self-Adaptive Optimized Link Prediction Based on Weak Ties Theory in Unweighted Network

A Balanced Modularity Maximization Link Prediction Model in Social Networks

A Bi-Scale Method Of Link Prediction

Multi-Tasking Link Prediction On Coupled Networks Via The Factor Graph Model

SCL-WTNS: A New Link Prediction Algorithm Based on Strength of Community Link and Weighted Two-Level Neighborhood Similarity

Exploiting neighbors' latent correlation for link prediction in complex network

A Community Detection and Graph-Neural-Network-Based Link Prediction Approach for Scientific Literature

An Improvement of Link Prediction by Combining Local Information and Betweenness

Toward Link Predictability Of Complex Networks

Link Prediction in Complex Networks Based on Communication Capacity and Local Paths

An Improved Link Prediction Algorithm Based on Degrees and Similarities of Nodes

Hidden Link Prediction Based on Node Centrality and Weak Ties

A Community Detection and Graph Neural Network Based Link Prediction Approach for Scientific Literature

Link Prediction in Complex Networks: a Local Naïve Bayes Model

Uncovering the network evolution mechanism by link prediction