Network embedding for link prediction: The pitfall and improvement

被引:29
作者
Cao, Ren-Meng [1 ]
Liu, Si-Yuan [1 ]
Xu, Xiao-Ke [1 ]
机构
[1] Dalian Minzu Univ, Coll Informat & Commun Engn, Dalian 116600, Peoples R China
基金
中国国家自然科学基金;
关键词
D O I
10.1063/1.5120724
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
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.
引用
收藏
页数:9
相关论文
共 33 条
[1]   Friendship Prediction and Homophily in Social Media [J].
Aiello, Luca Maria ;
Barrat, Alain ;
Schifanella, Rossano ;
Cattuto, Ciro ;
Markines, Benjamin ;
Menczer, Filippo .
ACM TRANSACTIONS ON THE WEB, 2012, 6 (02)
[2]  
[Anonymous], 1971, The Journal of mathematical sociology, DOI [10.1016/B978-0-12-442450-0.50012-2, DOI 10.1080/0022250X.1971.9989788]
[3]  
Bhagat S, 2011, SOCIAL NETWORK DATA ANALYTICS, P115
[4]   Link Prediction with Mutual Attention for Text-Attributed Networks [J].
Brochier, Robin ;
Guille, Adrien ;
Velcin, Julien .
COMPANION OF THE WORLD WIDE WEB CONFERENCE (WWW 2019 ), 2019, :283-284
[5]   A Comprehensive Survey of Graph Embedding: Problems, Techniques, and Applications [J].
Cai, HongYun ;
Zheng, Vincent W. ;
Chang, Kevin Chen-Chuan .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2018, 30 (09) :1616-1637
[6]  
Cao S, 2015, P 24 ACM INT C INF K, P891, DOI DOI 10.1145/2806416.2806512
[7]   The application of degree related clustering coefficient in estimating the link predictability and predicting missing links of networks [J].
Chen, Xing ;
Fang, Ling ;
Yang, Tinghong ;
Yang, Jian ;
Bao, Zerong ;
Wu, Duzhi ;
Zhao, Jing .
CHAOS, 2019, 29 (05)
[8]   Neural networks for link prediction in realistic biomedical graphs: a multi-dimensional evaluation of graph embedding-based approaches [J].
Crichton, Gamal ;
Guo, Yufan ;
Pyysalo, Sampo ;
Korhonen, Anna .
BMC BIOINFORMATICS, 2018, 19
[9]   A Survey on Network Embedding [J].
Cui, Peng ;
Wang, Xiao ;
Pei, Jian ;
Zhu, Wenwu .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2019, 31 (05) :833-852
[10]   Learning Structural Node Embeddings via Diffusion Wavelets [J].
Donnat, Claire ;
Zitnik, Marinka ;
Hallac, David ;
Leskovec, Jure .
KDD'18: PROCEEDINGS OF THE 24TH ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING, 2018, :1320-1329