计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (6A): 250700049-9.doi: 10.11896/jsjkx.250700049

• 大数据&数据科学 • 上一篇    下一篇

基于多特征融合与排序优化的图卷积神经网络关键节点识别

宋琳1, 王宇宁1, 石科仁2, 欧渊3   

  1. 1 西安建筑科技大学信息与控制工程学院 西安 710055
    2 中国人民解放军93756部队 天津 300000
    3 中国人民解放军32801部队 北京 100000
  • 出版日期:2026-06-16 发布日期:2026-06-12
  • 通讯作者: 欧渊(ouyunouyun55@163.com)
  • 作者简介:(songlin@xauat.edu.cn)

MFR-GCN:Key Node Identification via Multi-feature Fusion and Ranking Optimization in GraphConvolutional Networks

SONG Lin1, WANG Yuning1, SHI Keren2, OU Yuan3   

  1. 1 College of Information and Control Engineering,Xi'an University of Architecture and Technology,Xi'an 710055,China
    2 Unit 93756 of the PLA,Tianjin 300000,China
    3 Unit 32801 of the PLA,Beijing 100000,China
  • Published:2026-06-16 Online:2026-06-12
  • About author:SONG Lin,born in 1983,associate professor.His main research interests include pattern recognition,computer vision and information fusion.
    OU Yuan,born in 1984,associate researcher.His main research interests include information fusion and pattern recognition.

PDF (PC)

42

摘要： 针对现有复杂网络关键节点识别方法存在的特征融合浅层化、动态适应性不足及节点显著性差异弱化等问题,提出一种基于多特征融合与排序优化的图卷积神经网络(MFR-GCN)关键节点识别方法。该方法引入深度特征交互编码与可学习对比增强机制,通过层级自适应门控和条件化全局信息注入,实现关键节点的动态鲁棒检测。首先,从网络图中提取涵盖局部属性、全局属性、位置属性及随机游走属性等多维度的8个代表性特征,并结合节点嵌入构建特征向量;然后,将特征输入融合GCN与GAT的混合层进行深层特征学习,并利用跳跃连接整合多尺度信息;最后,设计包含排序损失、方差损失和聚类损失的增强版多组件损失函数进行模型训练与优化,并通过对比强化层在推理阶段放大关键节点分数差异,进一步区分关键节点。利用SIR传播模型在Cora,Email,C.elegans等真实网络数据集上进行实验验证。结果表明,相较于度中心性、介数中心性等传统方法,MFR-GCN识别出的关键节点在最终感染规模上平均提升显著(如在USairport网络上较次优方法提升约6.22%),展现出更优的全局传播潜力和适用性。

关键词: 关键节点识别, 图卷积神经网络, 多特征融合, 排序优化, 对比增强机制, 复杂网络

Abstract: To address the limitations of existing key node identification in the complex network,such as shallow feature fusion,insufficient dynamic adaptability,and weak differentiation of node significance,this paper proposes a graph convolutional network model based on multi-feature fusion and ranking optimization(MFR-GCN).The model innovatively incorporates deep feature interaction encoding and a learnable contrastive enhancement mechanism.It achieves dynamic and robust key node detection through hierarchical adaptive gating and conditional global information injection.Firstly,eight representative features spanning local attributes,global attributes,positional attributes,and random-walk properties(including the proposed LASPN centrality) are extracted from the network graph.These features are combined with node embeddings to construct feature vectors.Next,the vectors are fed into a hybrid layer integrating graph convolutional network(GCN) and graph attention network(GAT) for deep feature learning,while skip connections aggregate multi-scale information.Finally,an enhanced multi-component loss function－incorporating ranking loss,variance loss,and clustering loss—is designed for model training and optimization.During inference,a contrastive reinforcement layer amplifies the score differences of key nodes to further distinguish them.Validation experiments using the SIR propagation model are conducted on real-world datasets including Cora,Email,and C.elegans.Results demonstrate that compared to traditional methods like Degree Centrality and Betweenness Centrality,the key nodes identified by MFR-GCN achieve a significantly higher average final infection scale(such as exceeding the suboptimal method by approximately 6.22% on the USairport network).This highlights the model's superior global propagation potential and applicability.

Key words: Key node identification, Graph convolutional neural network, Multi-feature fusion, Ranking optimization, Contrast enhancement mechanism, Complex networks

中图分类号: 

  • TP330
[1] JU Y,ZHANG S,DING N,et al.Complex network clustering by a multi-objective evolutionary algorithm based on decomposition and membrane structure [J].SciRep:UK,2016,6:1.
[2] HAN Z M,WU Y,TAN X S,et al.Key node ranking in complex networks based on structural holes [J].Acta Physica Sinica,2015,64:058902.
[3] HUANG C L,FU R N,LI K Z.Partial switching topology iden-tification in dynamical networks under synchronization mechanisms [EB/OL].(2025-06-20) [2025-09-08] .https://link.cnki.net/urlid/37.1402.N.20250620.1055.002.
[4] WANG X,LI H.Key node identification algorithm in complex networks based on betweenness centrality entropy [J].Compu-ter & Digital Engineering,2024,52(3):677-680.
[5] VERNIZE G,GUEDES A L P,ALBINI L C P.Malicious nodes identification for complex network based on local views[J].The Computer Journal,2015,58(10):2476-2491.
[6] CARMI S,HAVLIN S,KIRKPATRICK S,et al.A model of Internet topology using k-shell decomposition[J].Proceedings of the National Academy of Sciences,2007,104(27):11150-11154.
[7] BRIN S,PAGE L.The anatomy of a large-scale hypertextualweb search engine[J].Computer Networks and ISDN Systems,1998,30(1/7):107-117.
[8] ZHU J F,CHEN D B,ZHOU T,et al.Survey on relative important node mining methods in network science [J].Journal of University of Electronic Science and Technology of China,2019,48:595.
[9] CHIANG W L,LIU X,SI S,et al.Cluster-gcn:An efficient algorithm for training deep and large graph convolutional networks[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2019:257-266.
[10] ZHONG L,GAO C,ZHANG Z,et al.Identifying influentialnodes in complex networks:A multiple attributes fusion method[C]//Active Media Technology:10th International Conference(AMT 2014).Springer,2014:11-22.
[11] KERMACK W O,MCKENDRICK A G.A contribution to the mathematical theory of epidemics[C]//Proceedings of the Royal Society of London.1927:700-721.
[12] SABIDUSSI G.The centrality index of a graph[J].Psycho-metrika,1966,31(4):581-603.
[13] SUN X Y,SHI Y C.Session recommendation model with graph neural network incorporating item influence [J].Journal of Computer Applications,2023,43(12):3689-3696.
[14] SUN X Y,SHI Y C.Session recommendation model with graph neural network incorporating item influence [J].Journal of Computer Applications,2023,43(12):3689-3696.
[15] RASHID Y,BHAT J I.OlapGN:A multi-layered graph convolution network-based model for locating influential nodes in graph networks[J].Knowledge-Based Systems,2024,283:111163.
[16] GUIMERAÀ R,DANON L,DÍAZ-GUILERA A,et al.Self-si-milar community structure in a network of human interactions[J].Physical Review E,2003,68(6):065103.
[17] CHRISTAKIS N A,FOWLER J H.The spread of obesity in a large social network over 32 years[J].New England Journal Mf medicine,2007,357(4):370-379.
[18] HAN Z M,CHEN Y,LI M Q,et al.An effective triangle-based model for measuring node influence in complex networks [J].Acta Physica Sinica,2016,65:168901.
[19] BELLINGERI M,BEVACQUA D,SCOTOGNELLA F,et al.A comparative analysis of link removal strategies in real complex weighted networks[J].Scientific Reports,2020,10(1):3911.
[20] ZHU C,WANG X,ZHU L.A novel method of evaluating key nodes in complex networks [J].Chaos Soliton Fract,2017,96:43.
[21] XIAO Y,CHEN Y,ZHANG H,et al.A new semi-local centrality for identifying influential nodes based on local average shortest path with extended neighborhood[J].Artificial Intelligence Review,2024,57(5):115.
[22] WANG D,CUI P,ZHU W.Structural deep network embedding[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2016:1225-1234.
[23] LU P,YANG J,LIU W.Identification of key nodes in complex networks by using a joint technique of nonnegative matrix factorization and regularization[J].Physical Communication,2024,65:102384.
[24] MCCALLUM A K,NIGAM K,RENNIE J,et al.Automating the construction of internet portals with machine learning [J].Information Retrieval,2000,3(2):127-163.
[25] GUIMERÀ R,DANON L,DÍAZ-GUILERA A,et al.Self-similar community structure in a network of human interactions [J].Physical Review E,2003,68(6):065103.
[26] WHITE J G,SOUTHGATE E,THOMSONJ N,et al.Thestructure of the nervous system of the nematode Caenorhabditis elegans [J].Philosophical Transactions of the Royal Society B,1986,314(1165):1-340.
[27] NEWMAN M E J.The structure of scientific collaboration networks [J].PNAS,2001,98(2):404-409.
[28] COLIZZA V,PASTOR-SATORRAS R,VESPIGNANI A.Re-action-diffusion processes and metapopulation models in heterogeneous networks [J].Nature Physics,2007,3(4):276-282.
[29] GUIMERÀ R,MOSSA S,TURTSCHI A,et al.The worldwide air transportation network:Anomalous centrality,community structure,and cities' global roles [J].PNAS,2005,102(22):7794-7799.
[30] YANG Y,WANG J F.Research on key node identification of complex network based on GCN [J].Journal of Sichuan University:Natural Science Edition,2023,60:032002.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发