[1]李格格,冶忠林,曹淑娟,等.一种近似图神经网络框架的无监督链路预测算法[J].郑州大学学报(工学版),2024,45(06):75-82.[doi:10.13705/j.issn.1671-6833.2024.03.011]
　LI Gege,YE Zhonglin,CAO Shujuan,et al.An Unsupervised Link Prediction Algorithm Based on an ApproximateGraph Neural Network Framework[J].Journal of Zhengzhou University (Engineering Science),2024,45(06):75-82.[doi:10.13705/j.issn.1671-6833.2024.03.011]
点击复制

一种近似图神经网络框架的无监督链路预测算法()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
45
期数:
2024年06期
页码:
75-82
栏目:
出版日期:
2024-09-25

文章信息/Info

Title:
An Unsupervised Link Prediction Algorithm Based on an ApproximateGraph Neural Network Framework
文章编号:
1671-6833(2024)06-0075-08
作者:
李格格1; 2;  冶忠林1; 2;  曹淑娟1; 2;  周 琳1; 2;  王雪力1; 2
1. 青海师范大学 计算机学院,青海 西宁 810008;2. 青海师范大学 藏语智能信息处理及应用国家重点实验室,青海,西宁 810008
Author(s):
LI Gege1; 2;  YE Zhonglin1; 2;  CAO Shujuan1; 2;  ZHOU Lin1; 2;  WANG Xueli1; 2
1. College of Computer, Qinghai Normal University, Xining 810008, China; 2. The State Key Laboratory of Tibetan Intelligent Information Processing and Application, Qinghai Normal University, Xining 810008, China
关键词:
矩阵分解;  向量优化;  图卷积神经网络;  相似度矩阵;  链路预测;  高阶近邻
Keywords:
matrix factorization;  vector optimization;  graph convolutional neural network;  similarity matrix;  link prediction;  high-order neighbors
分类号:
TP393. 0
DOI:
10.13705/j.issn.1671-6833.2024.03.011
文献标志码:
A
摘要:
对于无标签网络,由于基于图神经网络的链路预测方法使用其高效建模机制进行链路预测任务时性能较差,因此,提出了一种近似图神经网络框架的无监督链路预测算法( ALIP) ,旨在模拟图神经网络算法的高效建模机制和学习过程,解决网络节点标签缺失导致的建模不充分问题。 首先,参照 GCN 的输入层,融合网络的结构信息和节点属性;其次,使用矩阵分解替代 GCN 的隐藏层,模拟正向传播;再次,借鉴恒等映射和高阶近邻的思想实现向量转化和模型优化,从而得出网络节点表示向量,该过程模拟 GCN 的反向传播;最后,计算相似度矩阵,进行链路预测任务性能评测。 在 Citeseer 数据集、DBLP 数据集和 Cora 数据集上的实验结果表明:所提 ALIP 算法 AUC值最高为 98. 01%,其性能优于其他 23 种链路预测算法,证明了该算法的有效性和可行性,同时也为无标签的复杂网络链路预测任务提供了一种新的解决方案。
Abstract:
For unlabeled networks, the link prediction method based on graph neural networks had poor performance when using its efficient modeling mechanism for link prediction tasks. An unsupervised link prediction algorithm (ALIP) was proposed. It could approximate the graph neural network framework to simulate the efficientmodeling mechanism and learning process of graph neural network algorithms, and to solve the problem of insufficient modeling caused by missing network node labels. Firstly, referring to the input layer of GCN, the structuralinformation and node attributes of the network were fused. Secondly, matrix factorization is used to replace the hidden layer of GCN and simulate forward propagation. Then the ideas of identity mapping and vector optimization toachieve vector transformation and model optimization to obtain the network node representation vector, which wereused to simulate the back propagation of GCN. Finally, the similarity matrix for performance evaluation of link prediction tasks was calculated. On the Citeseer dataset, DBLP dataset and Cora dataset, the experimental resultsshowed that ALIP algorithm had a maximum AUC value of 98. 01%, and its performance was superior to the other23 link prediction algorithms. The effectiveness and feasibility of the algorithm, in this study provide a new solutionfor complex unlabeled network link prediction tasks.

参考文献/References:

[1] XIE X Q, LI Y J, ZHANG Z Q, et al. A joint link prediction method for social network[ C]∥International Conference of Young Computer Scientists, Engineers and Educators. Cham: Springer, 2015: 56-64.

[2] SCHAFER J L, GRAHAM J W. Missing data: our viewof the state of the art[ J] . Psychological Methods, 2002,7(2) : 147-177.
[3] KOSSINETS G. Effects of missing data in social networks[ J] . Social Networks, 2006, 28(3) : 247-268.
[4] LI M, MENG X M, ZHENG R Q, et al. Identification ofprotein complexes by using a spatial and temporal activeprotein interaction network[ J] . IEEE / ACM Transactionson Computational Biology and Bioinformatics, 2020, 17(3) : 817-827.
[5] DZAFERAGIC M, KAMINSKI N, MCBRIDE N, et al.A functional complexity framework for the analysis of telecommunication networks [ J ] . Journal of Complex Networks, 2018, 6(6) : 971-988.
[6] LIBEN-NOWELL D, KLEINBERG J. The link-predictionproblem for social networks[ J] . Journal of the AmericanSociety for Information Science and Technology, 2007, 58(7) : 1019-1031.
[7] LYU L Y, ZHOU T. Link prediction in complex networks: a survey [ J ] . Physica A: Statistical Mechanicsand Its Applications, 2011, 390(6) : 1150-1170.
[8] ZHOU T, LYU L Y, ZHANG Y C. Predicting missinglinks via local information [ J] . The European PhysicalJournal B, 2009, 71(4) : 623-630.
[9] JACCARD P. Etude de la distribution florale dans uneportion des Alpes et du Jura[ J] . Bulletin De La SocieteVaudoise Des Sciences Naturelles, 1901, 37( 142) : 547-579.
[10] SRENSEN T, SRENSEN T, BIERING-SRENSEN T,et al. A method of establishing group of equal amplitudein plant sociobiology based on similarity of species contentand its application to analyses of the vegetation on Danishcommons [ J] . Biologiske Skrifter, 1948, 5(4) : 1-34.
[11] 王富田, 张鹏,肖井华. 链路预测算法错边识别能力的评测[EB / OL]. (2015-12-30) [2023-05-15]. http:∥www. paper. edu. cn / releasepaper / content / 201512-1363.
WAMG F T, ZHANG P, XIAO J H. Evaluation the abilityof link prediction methods in the spurious link detection[EB / OL]. (2015-12-30) [2023-05-15]. http:∥www.paper. edu. cn / releasepaper / content / 201512-1363.
[12] KATZ L. A new status index derived from sociometric analysis[ J] . Psychometrika, 1953, 18(1) : 39-43.
[13] LEICHT E A, HOLME P, NEWMAN M E J. Vertexsimilarity in networks [ J] . Physical Review E, Statistical, Nonlinear, and Soft Matter Physics, 2006, 73(2) : 026120.
[14] FOUSS F, PIROTTE A, RENDERS J M, et al. Randomwalk computation of similarities between nodes of a graphwith application to collaborative recommendation [ J ] .IEEE Transactions on Knowledge and Data Engineering,2007, 19(3) : 355-369.
[15] KOVÁCS I A, LUCK K, SPIROHN K, et al. Networkbased prediction of protein interactions[ J] . Nature Communications, 2019, 10: 1240.
[16] 廖亮, 张恒锋. 基于支持向量机的机会网络链路预测[ J] . 信息通信, 2018, 31(9) : 23-25.
LIAO L, ZHANG H F. Link prediction of opportunisticnetwork based on support vector machine[ J] . Information& Communications, 2018, 31(9) : 23-25.
[17] 杨妮亚, 彭涛, 刘露. 基于聚类和决策树的链路预测方法[ J] . 计 算 机 研 究 与 发 展, 2017, 54 ( 8) : 1795-1803.
YANG N Y, PENG T, LIU L. Link prediction methodbased on clustering and decision tree [ J ] . Journal ofComputer Research and Development, 2017, 54 ( 8 ) :1795-1803.
[18] KIPF T N, WELLING M. Semi-supervised classificationwith graph convolutional networks [ EB / OL]. ( 2017 - 02 -22)[2023-05-11]. https:∥arxiv. org / abs/ 1609. 02907.
[19] LEI K, QIN M, BAI B, et al. GCN-GAN: a non-lineartemporal link prediction model for weighted dynamic networks[C]∥IEEE INFOCOM 2019 - IEEE Conference onComputer Communications. Piscataway: IEEE, 2019:388-396.
[20] LI G H, MÜLLER M, THABET A, et al. DeepGCNs:can GCNs go as deep as CNNs? [ C]∥2019 IEEE / CVFInternational Conference on Computer Vision ( ICCV ) .Piscataway: IEEE, 2019: 9266-9275.
[21] CHEN M, WEI Z W, HUANG Z F, et al. Simple anddeep graph convolutional networks [ EB / OL] ( 2020 - 07 -04)[2023-05-11]. https:∥arxiv. org / abs/ 2007. 02133.
[22] YANG C, SUN M S, LIU Z Y, et al. Fast network embedding enhancement via high order proximity approximation[ C] ∥Proceedings of the Twenty-Sixth InternationalJoint Conference on Artificial Intelligence. Melbourne:International Joint Conferences on Artificial IntelligenceOrganization, 2017: 3894-3900.
[23] LIU Z, ZHANG Q M, LYU L Y, et al. Link predictionin complex networks: a local naïve Bayes model[ J] . EPL(Europhysics Letters) , 2011, 96(4) : 48007.
[24] KLEIN D J, RANDI? M. Resistance distance[ J] . Journal of Mathematical Chemistry, 1993, 12(1) : 81-95.
[25] CHEBOTAREV P, SHAMIS E. The matrix-forest theorem and measuring relations in small social groups [ EB /OL] . ( 2006 - 02 - 04) [ 2023 - 06 - 01] . https:∥arxiv.org / abs/ math / 0602070.
[26] 冶忠林, 曹蓉, 赵 海 兴, 等. 基 于 矩 阵 分 解 的 DeepWalk 链路预测算法[ J] . 计算机应用研究, 2020, 37(2) : 424-429, 442.
YE Z L, CAO R, ZHAO H X, et al. Link predictionbased on matrix factorization for DeepWalk[ J] . Application Research of Computers, 2020, 37 ( 2 ) : 424 -429, 442.
[27] 曹蓉, 赵海兴, 冶忠林. 基于网络节点文本增强的链路预测算法[ J] . 计算机应用与软件, 2019, 36( 3) :227-235, 242.
CAO R, ZHAO H X, YE Z L. Link prediction algorithmbased on text enhanced of network nodes [ J] . ComputerApplications and Software, 2019, 36 ( 3 ) : 227 -235, 242.
[28] 王曙燕, 巩婧怡. 融合节点标签与强弱关系的链路预测算法[ J] . 计算机工程与应用, 2022, 58 ( 18) : 71-77.
WANG S Y, GONG J Y. Link prediction algorithm fusingComputer Engineering and Applications, 2022, 58(18) : 71-77.

更新日期/Last Update: 2024-09-29