一种预测未知节点的融合影响力最大化<br />的知识可迁移GNN模型

PDF(4676 KB)

中文信息学报 ›› 2025, Vol. 39 ›› Issue (2) : 89-99,110.

情感分析与社会计算

一种预测未知节点的融合影响力最大化
的知识可迁移GNN模型

曾志林¹,张超群^1,2,吴国富¹,汤卫东¹,李灏然¹,李婉秋¹

作者信息 +

A Knowledge Transferable GNN Model Integrating Influence Maximization for Predicting Unknown Nodes

ZENG Zhilin¹, ZHANG Chaoqun^1,2, WU Guofu¹, TANG Weidong¹, LI Haoran¹, LI Wanqiu¹

Author information +

History +

摘要

在社交网络中,大多数节点的数据不完整,已有的方法对这些节点的预测效率较低。鉴于此,该文提出一种融合影响力最大化的知识可迁移图神络网络(Graph Neural Network,GNN)模型VRKTGNN,其是对预测社交网络未知节点的KTGNN模型的改进。VRKTGNN根据用户的关注去构建一个图结构数据,由改进的投票排名算法VoteRank++选出图数据中影响力最大的节点对未知节点进行知识迁移,通过KTGNN利用影响力最大的节点将未知节点的信息进行完善或者补全,进而预测出大多数未知节点的一个关注重点。在五个数据集上的实验结果表明,VRKTGNN总体明显优于十个对比模型。具体来说,与最优的对比模型KTGNN相比,VRKTGNN在Github-web数据集上性能非常接近,而在Twitch-DE、Tolokers、Twitter、Twitch-EN数据集上的F₁值分别提升5.73%、2.9%、2.86%和1.83%。这些结果均表明,该文新提出的模型鲁棒性更强,能够利用影响力最大的节点对社交网络中的未知节点进行有效预测,且对复杂网络更具优势。

Abstract

In social networks, most nodes have incomplete data, and existing methods are less efficient in predicting these nodes. This paper proposes a knowledge transferable Graph Neural Network (GNN) model called VRKTGNN that integrates influence maximization, which is an improvement of the KTGNN model for predicting unknown nodes in social networks. VRKTGNN constructs a graph structure data based on users concerns, and uses the improved vote ranking algorithm VoteRank++ to select the most influential nodes from the graph data. Furthermore, KTGNN uses the most influential nodes to refine or complement the information of unknown nodes, thus predicting the attentional focus of most unknown nodes. Experiments on five well-known datasets show that VRKTGNN significantly outperforms the ten competitors, with 5.73%, 2.9%, 2.86% and 1.83% increase compared with KTGNN in F₁-Score the Twitch-DE, Tolokers, Twitter and Twitch-EN datasets, respectively. These results suggest that the proposed model performs rabustness, which can effectioely predict the unknown nodes using the most influential nodes in social networks and has more advantages for comples networks.

导出引用

曾志林,张超群,吴国富,汤卫东,李灏然,李婉秋. 一种预测未知节点的融合影响力最大化
的知识可迁移GNN模型. 中文信息学报. 2025, 39(2): 89-99,110

ZENG Zhilin, ZHANG Chaoqun, WU Guofu, TANG Weidong, LI Haoran, LI Wanqiu. A Knowledge Transferable GNN Model Integrating Influence Maximization for Predicting Unknown Nodes. Journal of Chinese Information Processing. 2025, 39(2): 89-99,110

参考文献

[1] FRANCO S, MARCO G, AH C T, et al. The graph neural network model[J]. IEEE Trans on Neural Networks, 2009,20(1): 61-80.
[2] ZHOU J, CUI G, HUANG C, et al. Graph neural networks: A review of methods and applications[J]. AI Open, 2020, 1: 57-81.
[3] ZHANG C, SONG D, HUANG C, et al. Heterogeneous graph neural network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM Press, 2019: 793-803.
[4] BI W, XU B, SUN X, et al. Predicting the silent majority on graphs: Knowledge transferable graph neural network[C]//Proceedings of the ACM Web Conference 2023. New York: ACM Press, 2023: 274-285.
[5] MAKSIM K, LAZAROS K G, SHLOMO H, et al. Identification of influential spreaders in complex networks[J]. Nature Physics, 2010(6): 888-893.
[6] AHMAD Z, AMIR S, KEYHAN K. Influence maximization in social networks based on TOPSIS[J]. Expert Systems with Applications, 2018, 108: 96-107.
[7] LIU P, LI L, FANG S, et al. Identifying influential nodes in social networks: A voting approach[J]. Chaos, Solitons & Fractals, 2021, 152, 111309.
[8] THOMAS K, MAX W. Semi-supervised classification with graph convolutional networks[C]//Proceedings of 5th International Conference on Learning Representations, 2016.
[9] PETAR V, GUILLEM C, ARANTXA C, et al. Graph attention networks[C]//Proceedings of 6th International Conference on Learning Representations, 2017.
[10] WILLIAM L H, REX Y, JURE L. Inductive representation learning on large graphs[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM Press, 2017: 1025-1035.
[11] LONG M, CAO Y, WANG J, et al. Learning transferable features with deep adaptation networks[C]//Proceedings of the 32nd International Conference on International Conference on Machine Learning. New York: ACM Press, 2015: 97-105.
[12] LONG M, CAO Z, WANG J, et al. Conditional adversarial domain adaptation[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems. New York: ACM Press, 2018: 1647-1657.
[13] DAI Q, Wu X M, XIAO J, et al. Graph transfer learning via adversarial domain adaptation with graph convolution[J]. IEEE Trans on Knowledge and Data Engineering, 2023, 35(5): 4908-4922.
[14] SHEN X, PAN S, CHOI K Z, et al. Domain-adaptive message passing graph neural network[J]. Neural Networks, 2023, 164: 439-454.
[15] DAVID K, JON K, VA T. Maximizing the spread of influence through a social network[C]//Proceedings of the 9th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM Press, 2003: 137-146.
[16] CHEN D, LU L, SHANG M S, et al. Identifying influential nodes in complex networks[J]. Physica A: Statistical Mechanics and its Applications, 2012, 391(4): 1777-1787.
[17] ZHANG X, ZHU J, WANG Q, et al. Identifying influential nodes in complex networks with community structure[J]. Knowledge-Based Systems, 2013, 42: 74-84.
[18] QIU J, TANG J, MA H, et al. DeepInf: Social influence prediction with deep learning[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM Press, 2018: 2110-2119.
[19] KOU J, JIA P, LIU J. et al. Identify influential nodes in social networks with graph multi-head attention regression model[J]. Neurocomputing, 2023, 530: 23-36.
[20] LI W, HU Y, JINAG C, et al. ABEM: An adaptive agent-based evolutionary approach for influence maximization in dynamic social networks[J]. Applied Soft Computing, 2023, 136: 110062.
[21] 刘小洋,唐婷,何道兵. 融合社交网络用户自身属性的信息传播数学建模与舆情演化分析[J].中文信息学报, 2019, 33(9): 115-122.
[22] 张陶, 于炯, 廖彬, 等. 基于图嵌入与支持向量机的社交网络节点分类方法[J]. 计算机应用研究, 2021, 38(9): 2646-2650,2661.
[23] BENEDEK R, CARL A, RIK S, et al. Multi-scale attributed node embedding[J]. Journal of Complex Networks, 2021, 9(1): 1-22.
[24] OLEG P, DENIS K, MICHAEL D. et al. A critical look at the evaluation of GNNs under heterophily: Are we really making progress?[C]//Proceedings of 12th International Conference on Learning Representations, 2023.
[25] XU K, LI C, TIAN Y, et al. Representation learning on graphs with jumping knowledge networks[C]//Proceedings of the 35th International Conference on Machine Learning, 2018: 5453-5462.
[26] JOHANNES G, ALEKSANDAR B, STEHPHAN G. Predict then propagate: Graph neural networks meet personalized pagerank[C]//Proceedings of 7th International Conference on Learning Representations, 2018.
[27] SHAKED B, URI A, ERAN Y. How attentive are graph attention networks?[C]//Proceedings of 10th International Conference on Learning Representations, 2021.[28] LIU M, GAO H, JI S. Towards deeper graph neural networks[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM Press, 2020: 338-348.
[29] CHEN M, WEI Z, HUANG Z, et al. Simple and deep graph convolutional networks[C]//Proceedings of the 37th International Conference on Machine Learning. New York: ACM Press, 2020: 1725-1735.
[30] SONG Y, WANG D. Learning on graphs with out-of-distribution nodes[C]//Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: Press, 2022: 1635-1645.