计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (7): 213-221.doi: 10.11896/jsjkx.250700055

• 数据库 & 大数据 & 数据科学 • 上一篇    下一篇

基于因果子图学习的级联流行度预测

李锴钜, 尹晨阳, 程章桃, 刘学婷, 周帆   

  1. 电子科技大学信息与软件工程学院 成都 610054
  • 收稿日期:2025-07-10 修回日期:2025-09-28 出版日期:2026-07-15 发布日期:2026-07-10
  • 通讯作者: 周帆(fan.zhou@uestc.edu.cn)
  • 作者简介:(lkj2015@outlook.com)
  • 基金资助:
    :国家自然科学基金(62176043,62072077,U22A2097)

Causal Subgraph Learning for Cascade Popularity Prediction

LI Kaiju, YIN Chenyang, CHENG Zhangtao, LIU Xueting, ZHOU Fan   

  1. School of Information and Software Engineering,University of Electronic Science and Technology of China,Chengdu 610054,China
  • Received:2025-07-10 Revised:2025-09-28 Published:2026-07-15 Online:2026-07-10
  • About author:LI Kaiju,born in 1989,master,research associate.His main research interests include event prediction,social network data mining and recommender system.
    ZHOU Fan,born in 1981,Ph.D,professor,is a member of CCF(No.C3844M).His main research interests include machine learning,spatio-temporal data mining,data mining and know-ledge discovery.
  • Supported by:
    National Natural Science Foundation of China(62176043,62072077,U22A2097).

PDF (PC)

15

摘要： 信息级联的流行度预测对于理解信息传播规律和抑制不实信息扩散至关重要。现有深度学习方法虽在准确性上有所提升,但在消解混淆因素、挖掘级联结构与流行度间的深层因果关系方面仍存在不足,影响了预测的可靠性与可解释性。为应对此挑战,提出了一种新颖的基于因果子图学习的级联流行度预测模型——因果感知级联模型(Causal-aware Cascade Model,CauCas)。CauCas设计图数据增强对原始级联图施加干预,并分别对原始图和增广图进行多层级联表示编码,通过特定的层特征选择与加权融合策略,得到原始图和增广图各自的图级别表示,并使用自适应实例归一化技术学习到对干预更不敏感、更可能反映因果关系的鲁棒特征。最终,融合后的特征表示通过一个多层感知机进行流行度预测。在Twitter,Weibo和APS这3个公开数据集上的实验结果表明,CauCas模型性能优越,在不同类型数据集和预测窗口下均达到了最优性能。

关键词: 因果子图, 信息级联, 流行度预测, 因果推断, 图神经网络, 注意力机制

Abstract: Information cascade popularity prediction is critical for understanding the dynamics of information dissemination and mitigating the spread of misinformation.Although existing deep learning methods have achieved improvements in predictive accuracy,they still exhibit limitations in disentangling confounding factors and uncovering the deep causal relationships between cascade structures and popularity,which undermines both the reliability and interpretability of predictions.To address these challenges,this paper proposes a novel cascade popularity prediction model based on causal subgraph learning,named causal-aware cascade model(CauCas).CauCas introduces graph data augmentation to impose interventions on the original cascade graphs and encodes multi-level cascade representations for both the original and augmented graphs.Through specialized layer-wise feature selection and weighted fusion strategies,the model derives graph-level representations for each graph,and leverages adaptive instance normalization to learn robust features that are less sensitive to interventions and more likely to reflect causal relationships.Finally,the fused feature representations are fed into a multilayer perceptron to perform popularity prediction.Experimental results on Twitter,Weibo,and APS three public datasets demonstrate that CauCas achieves superior performance,consistently outperforming state-of-the-art methods across diverse datasets and prediction windows.

Key words: Causal subgraph, Information cascade, Popularity prediction, Causal inference, Graph neural networks, Attention mechanism

中图分类号: 

  • TP391
[1]ROBERTSON C E,PRÖLLOCHS N,SCHWARZENEGGERK,et al.Negativity drives online news consumption[J].Nature Human Behaviour,2023,7(5):812-822.
[2]ZHOU F,XU X,TRAJCEVSKI G,et al.A survey of information cascade analysis:Models,predictions,and recent advances[J].ACM Computing Surveys,2021,54(2):1-36.
[3]LIU L,CHEN J,CHENG Z,et al.Towards Trustworthy Rumor Detection with Interpretable Graph Structural Learning[C]//Proceedings of the 32nd ACM International Conference on Information and Knowledge Management.2023:4089-4093.
[4]TATAR A,DE AMORIM M D,FDIDA S,et al.A survey on predicting the popularity of web content[J].Journal of Internet Services and Applications,2014,5:1-20.
[5]MISHRA S,RIZOIU M A,XIE L.Feature driven and pointprocess approaches for popularity prediction[C]//Proceedings of the 25th ACM International on Conference on Information and Knowledge Management.2016:1069-1078.
[6]CHENG J,ADAMIC L,DOWP A,et al.Can cascades be predicted?[C]//Proceedings of the 23rd International Conference on World Wide Web.2014:925-936.
[7]ZHAO Q,ERDOGDU M A,HE H Y,et al.Seismic:A self-exci-ting point process model for predicting tweet popularity[C]//Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2015:1513-1522.
[8]RIZOIU M A,MISHRA S,KONG Q,et al.SIR-Hawkes:Lin-king epidemic models and Hawkes processes to model diffusions in finite populations[C]//Proceedings of the 2018 World Wide Web Conference.2018:419-428.
[9]XU X,ZHOU F,ZHANG K,et al.Casflow:Exploring hierarchical structures and propagation uncertainty for cascade prediction[J].IEEE Transactions on Knowledge and Data Engineering,2021,35(4):3484-3499.
[10]ZHANG Z,XIE X,ZHANG Y,et al.Hier Cas:Hierarchical Temporal Graph Attention Networks for Popularity Prediction in Information Cascades[C]//2024 International Joint Confe-rence on Neural Networks(IJCNN).IEEE,2024:1-8.
[11]HUANG X,BELONGIE S.Arbitrary style transfer in real-timewith adaptive instance normalization[C]//Proceedings of the IEEE International Conference on Computer Vision.2017:1501-1510.
[12]SZABO G,HUBERMANB A.Predicting the popularity of on-line content[J].Communications of the ACM,2010,53(8):80-88.
[13]ROMERO D,TAN C,UGANDER J.On the interplay between social and topical structure[C]//Proceedings of the Internatio-nal AAAI Conference on Web and Social Media.2013:516-525.
[14]BAO P,SHEN H W,HUANG J,et al.Popularity prediction in microblogging network:a case study on sina weibo[C]//Proceedings of the 22nd International Conference on World Wide Web.2013:177-178.
[15]TSUR O,RAPPOPORTA.What's in a hashtag? Content based prediction of the spread of ideas in microblogging communities[C]//Proceedings of the Fifth ACM International Conference on Web Search and Data Mining.2012:643-652.
[16]SHEN H,WANG D,SONG C,et al.Modeling and predictingpopularity dynamics via reinforced poisson processes[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2014.
[17]BAO P,SHEN H W,JIN X,et al.Modeling and predicting popularity dynamics of microblogs using self-excited hawkes processes[C]//Proceedings of the 24th International Conference on World Wide Web.2015:9-10.
[18]YU L,CUI P,WANG F,et al.From micro to macro:Uncovering and predicting information cascading process with behavioral dynamics[C]//2015 IEEE International Conference on Data Mi-ning.IEEE,2015:559-568.
[19]GAO X,CAO Z,LI S,et al.Taxonomy and evaluation for microblog popularity prediction[J].ACM Transactions on Know-ledge Discovery from Data,2019,13(2):1-40.
[20]LU X,JI S,YU L,et al.Continuous-time graph learning for cascade popularity prediction[C]//Proceedings of the Thirty-Se-cond International Joint Conference on Artificial Intelligence.2023:2224-2232.
[21]YANG C,WANG H,TANG J,et al.Full-scale information diffusion prediction with reinforced recurrent networks[J].IEEE Transactions on Neural Networks and Learning Systems,2021,34(5):2271-2283.
[22]CHEN X,ZHOU F,ZHANG K,et al.Information diffusion prediction via recurrent cascades convolution[C]//2019 IEEE 35th International Conference on Data Engineering(ICDE).IEEE,2019:770-781.
[23]NGUYEN T,DO K,NGUYEN D T,et al.Causal inference via style transfer for out-of-distribution generalisation[C]//Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining.2023:1746-1757.
[24]WU P,LI H,DENG Y,et al.On the opportunity of causal lear-ning in recommendation systems:Foundation,estimation,prediction and challenges[C]//Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence.2022:5646-5653
[25]LEE N,YOON K,NA G S,et al.Shift-robust molecular relational learning with causal substructure[C]//Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining.2023:1200-1212.
[26]JEONG H,TIAN J,BAREINBOIM E.Finding and listing front-door adjustment sets[J].Advances in Neural Information Processing Systems,2022,35:33173-33185.
[27]SHAH A,SHANMUGAM K,KOCAOGLU M.Front-door adjustment beyond markov equivalence with limited graph know-ledge[C]//Advances in Neural Information Processing Systems.2024.
[28]ZHAO T,JIN W,LIU Y,et al.Graph Data Augmentation for Graph Machine Learning:A Survey[J].arXiv.2202.08871,2022.
[29]WENG L,MENCZER F,AHN Y Y.Virality prediction andcommunity structure in social networks[J].Scientific Reports,2013,3(1):1-6.
[30]CAO Q,SHEN H,CEN K,et al.Deephawkes:Bridging the gap between prediction and understanding of information cascades[C]//Proceedings of the 2017 ACM on Conference on Information and Knowledge Management.2017:1149-1158.
[31]PINTO H,ALMEIDA J M,GONÇALVES M A.Using early view patterns to predict the popularity of youtube videos[C]//Proceedings of the Sixth ACM International Conference on Web Search and Data Mining.2013:365-374.
[32]CHENG G,YAN X,GAO S,et al.Cas Sampling:Exploring Efficient Cascade Graph Learning for Popularity Prediction[C]//Joint European Conference on Machine Learning and Knowledge Discovery in Databases.Cham:Springer,2023:70-86.
[33]ZHOU M,LIN Y,LIU G,et al.Modeling Personalized Retwee-ting Behaviors for Multi-Stage Cascade Popularity Prediction[C]//Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence.2024:2598-2606.
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