Please wait a minute...
 首页  期刊介绍 期刊订阅 联系我们 横山亮次奖 百年刊庆
最新录用  |  预出版  |  当期目录  |  过刊浏览  |  阅读排行  |  下载排行  |  引用排行  |  横山亮次奖  |  百年刊庆
清华大学学报(自然科学版)  2024, Vol. 64 Issue (4): 688-699    DOI: 10.16511/j.cnki.qhdxxb.2024.22.006
  信息科学与技术 本期目录 | 过刊浏览 | 高级检索 |
张雪芹1,2, 刘岗1, 王智能1, 罗飞1, 吴建华2
1. 华东理工大学 信息科学与工程学院, 上海 200237;
2. 上海市计算机软件评测重点实验室, 上海 201112
Microscopic diffusion prediction based on multifeature fusion and deep learning
ZHANG Xueqin1,2, LIU Gang1, WANG Zhineng1, LUO Fei1, WU Jianhua2
1. School of Information Science and Engineering, East China University of Science and Technology, Shanghai 200237, China;
2. Shanghai Key Laboratory of Computer Software Evaluating and Testing, Shanghai 201112, China
全文: PDF(5904 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 准确地预测社交网络中的信息扩散节点可以对谣言、 计算机病毒等不良信息的传播以及信息泄露做到早检测、 早溯源和早抑制。 为了提高微观扩散预测精度, 该文提出了一个基于多特征融合和深度学习的微观信息扩散预测通用框架(MFFDLP)。 为了获取信息扩散的时序特征, 基于信息扩散序列和社交网络图, 采用门控循环神经网络提取局部时序特征和全局时序特征, 并融合形成信息扩散序列表征; 为了获取用户交互行为和兴趣爱好的动态表示, 根据历史信息构建信息扩散图, 使用级联图注意力网络提取信息扩散子图中节点特征和边特征, 并通过嵌入查找, 融合形成当前信息扩散序列中相应节点的动态扩散表征; 使用双多头注意力机制, 进一步捕获静态和动态扩散特征的上下文信息, 实现了高精度微观扩散预测。 在3个公共数据集上的对比实验结果表明: 所提方法优于对比方法, 在微观扩散预测的精度上最高提高了9.98%。
E-mail Alert
关键词 社交网络微观扩散预测循环神经网络图注意力网络多头注意力机制    
Abstract:[Objective] Deep learning methods have been widely employed to enhance microscopic diffusion prediction in social networks. However, the existing methods have the problem of insufficient extraction of features in the information dissemination process. For example, these methods do not consider the impact of the propagation chain of the most recently infected nodes on the subsequent propagation of the message or the impact of changes in the neighboring nodes on the propagation path of the message. Therefore, the prediction accuracy is not high. To solve the above problems, describe the information diffusion process from multiple perspectives, and discover more hidden features, this paper proposes a microscopic diffusion prediction framework — multifeature fusion and deep learning for prediction (MFFDLP). [Methods] The microscopic diffusion prediction framework is divided into three main parts: extracting the static features from the network topology and the information diffusion sequence, capturing dynamic diffusion characteristics from the information diffusion graph, and predicting the next infected node. (1) First, node embedding and node structure context are extracted from historical friendship graphs and information diffusion sequences. The gate recurrent unit (GRU) is applied to mine the deep global temporal features from the connected vectors. To further enhance the role of the recently infected node, GRU is used to mine the local temporal features from the structure context of the node. These two features are fused to form the information diffusion sequence features. (2) Capture dynamic diffusion characteristics from the information diffusion graph. These features represent changes in users' interaction or interest. An information diffusion graph is built based on the historical information diffusion sequence. The diffusion graph is then divided into subgraphs in chronological order. A graph attention network is applied to capture the node features from each subgraph, and the edge features are aggregated from the node features. Using an embedding lookup method and fusing the nodes and their edge features, the dynamic diffusion characteristics of the users in an information diffusion sequence are obtained. (3) Predict the next infected node. To further analyze the context interaction within the diffusion sequences, a dual multihead self-attention mechanism is applied to separately capture the contextual information from information diffusion sequence features and node dynamic diffusion characteristics. Then, a fully connected layer and Softmax are used to predict the next infected node. Finally, experiments on three real networks show that the proposed method outperforms the state-of-the-art models. The experimental results demonstrate the unique advantages of MFFDLP for microscopic diffusion prediction. [Results] Comparative experimental results on three public datasets show that the proposed method outperforms the comparative methods by up to 9.98% in the accuracy of microscopic diffusion prediction. [Conclusions] This method comprehensively combines the friendship graph, information diffusion sequence, and diffusion graph. Multiple deep learning models are used to extract multiple features from static and dynamic perspectives. Comparative experiments on multiple datasets demonstrate that MFFDLP can mine and fuse multiple features more effectively, thus improving the prediction accuracy of information diffusion.
Key wordssocial network    microscopic diffusion prediction    recurrent neural network    graph attention network    multihead attention mechanism
收稿日期: 2023-10-25      出版日期: 2024-03-27
作者简介: 张雪芹(1972—),女,教授。
张雪芹, 刘岗, 王智能, 罗飞, 吴建华. 基于多特征融合和深度学习的微观扩散预测[J]. 清华大学学报(自然科学版), 2024, 64(4): 688-699.
ZHANG Xueqin, LIU Gang, WANG Zhineng, LUO Fei, WU Jianhua. Microscopic diffusion prediction based on multifeature fusion and deep learning. Journal of Tsinghua University(Science and Technology), 2024, 64(4): 688-699.
链接本文:  或
[1] Statista. Social media: Statistic & facts[R/OL]. (2023-08-31)[2023-08-31].
[2] MEEL P, VISHWAKARMA D K. Fake news, rumor, information pollution in social media and web: A contemporary survey of state-of-the-arts, challenges and opportunities[J]. Expert Systems with Applications, 2020, 153: 112986.
[3] GRINBERG N, JOSEPH K, FRIEDLAND L, et al. Fake news on Twitter during the 2016 U.S. presidential election[J]. Science, 2019, 363(6425): 374-378.
[4] 张志扬, 张凤荔, 谭琪, 等. 基于深度学习的信息级联预测方法综述[J]. 计算机科学, 2020, 47(7): 141-153. ZHANG Z Y, ZHANG F L, TAN Q, et al. Review of information cascade prediction methods based on deep learning[J]. Computer Science, 2020, 47(7): 141-153. (in Chinese)
[5] MATSUBARA Y, SAKURAI Y, PRAKASH B A, et al. Rise and fall patterns of information diffusion: Model and implications[C]//Proceedings of the 18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Beijing, China: Association for Computing Machinery, 2012: 6-14.
[6] YU L Y, CUI P, WANG F, et al. From micro to macro: Uncovering and predicting information cascading process with behavioral dynamics[C]//Proceedings of 2015 IEEE International Conference on Data Mining. Atlantic, USA: IEEE, 2015: 559-568.
[7] CAO X Y, CHEN Y, JIANG C X, et al. Evolutionary information diffusion over heterogeneous social networks[J]. IEEE Transactions on Signal and Information Processing over Networks, 2016, 2(4): 595-610.
[8] ZHAO J H, ZHAO J L, FENG J. Information diffusion prediction based on cascade sequences and social topology[J]. Computers and Electrical Engineering, 2023, 109: 108782.
[9] FENG S S, ZHAO K Q, FANG L T, et al. H-Diffu: Hyperbolic representations for information diffusion prediction[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(9): 8784-8798.
[10] 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, 2023, 34(5): 2271-2283.
[11] WANG Z T, CHEN C Y, LI W J. A sequential neural information diffusion model with structure attention[C]//Proceedings of the 27th ACM International Conference on Information and Knowledge Management. Torino, Italy: Association for Computing Machinery, 2018: 1795-1798.
[12] HUANG N B, ZHOU G, ZHANG M L, et al. MC-RGCN: A multi-channel recurrent graph convolutional network to learn high-order social relations for diffusion prediction[C]//Proceedings of 2021 IEEE International Conference on Data Mining. Auckland, New Zealand: IEEE, 2021: 1108-1113.
[13] FATEMI B, MOLAEI S, PAN S R, et al. GCNFusion: An efficient graph convolutional network based model for information diffusion[J]. Expert Systems with Applications, 2022, 202: 117053.
[14] YUAN C Y, LI J C, ZHOU W, et al. DyHGCN: A dynamic heterogeneous graph convolutional network to learn users' dynamic preferences for information diffusion prediction[C]//Proceedings of European Conference Machine Learning and Knowledge Discovery in Databases. Ghent, Belgium: Springer-Verlag, 2020: 347-363.
[15] WANG R J, HUANG Z J, LIU S Z, et al. DyDiff-VAE: A dynamic variational framework for information diffusion prediction[C/OL]//Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. Virtual Event. New York, USA: Association for Computing Machinery, 2021: 163-172.
[16] CAO Z M, HAN K, ZHU J H. Information diffusion prediction via dynamic graph neural networks[C]//Proceedings of the IEEE 24th International Conference on Computer Supported Cooperative Work in Design. Dalian, China: IEEE, 2021: 1099-1104.
[17] SUN L, BAO Y, ZHANG X B, et al. MS-HGAT: Memory-enhanced sequential hypergraph attention network for information diffusion prediction[C/OL]//Proceedings of the 36th AAAI Conference on Artificial Intelligence. 34th Conference on Innovative Applications of Artificial Intelligence. The 12th Symposium on Educational Advances in Artificial Intelligence. Virtual Event. Washington DC, USA: AAAI Press, 2022: 4156-4164.
[18] ZHAO Q H, ZHANG Y Z, FENG X D. Predicting information diffusion via deep temporal convolutional networks[J]. Information Systems, 2022, 108: 102045.
[19] MIYAZAWA H, MURATA T. Graph convolutional network with time-based mini-batch for information diffusion prediction[C]//Proceedings of the 9th International Conference on Complex Networks and Their Applications. Cham, Germany: Springer, 2021: 53-65.
[20] MIKOLOV T, KARAFIáT M, BURGET L, et al. Recurrent neural network based language model[C]//Proceedings of the 11th Annual Conference of the International Speech Communication Association. Makuhari, Japan: DBLP, 2010: 1045-1048.
[21] CHUNG J, GULCEHRE C, CHO K H, et al. Empirical evaluation of gated recurrent neural networks on sequence modeling[C]//NIPS 2014 Workshop on Deep Learning. Montreal, Canada: NIPS, 2014.
[22] VELI AČU KOVI AĆU P, CASANOVA A, LIÒ P, et al. Graph attention networks[C]//Proceedings of the 6th International Conference on Learning Representations. Vancouver, Canada: ICLR, 2018.
[23] PEROZZI B, AL-RFOU R, SKIENA S. DeepWalk: Online learning of social representations[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: Association for Computing Machinery, 2014: 701-710.
[24] MIKOLOV T, CHEN K, CORRADO G, et al. Efficient estimation of word representations in vector space[C]//Proceedings of the 1st International Conference on Learning Representations. Scottsdale, USA: ICLR, 2013.
[25] KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks[C]//Proceedings of the 5th International Conference on Learning Representations. Toulon, France: ICLR, 2017: 1-14.
[26] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: Curran Associates, 2017: 6000-6010.
[27] HODAS N O, LERMAN K. The simple rules of social contagion[J]. Scientific Reports, 2014, 4(1): 4343.
[28] ZHONG E H, FAN W, WANG J W, et al. ComSoc: Adaptive transfer of user behaviors over composite social network[C]//Proceedings of the 18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Beijing, China: IEEE, 2012: 696-704.
[29] LESKOVEC J, BACKSTROM L, KLEINBERG J. Meme-tracking and the dynamics of the news cycle[C]//Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Paris, France: Association for Computing Machinery, 2009: 497-506.
[30] PASZKE A, GROSS S, MASSA F, et al. PyTorch: An imperative style, high-performance deep learning library[C]//Proceedings of the 33rd International Conference on Neural Information Processing Systems. Vancouver, Canada: Curran Associates, 2019: 721.
[1] 张名芳, 李桂林, 吴初娜, 王力, 佟良昊. 基于轻量型空间特征编码网络的驾驶人注视区域估计算法[J]. 清华大学学报(自然科学版), 2024, 64(1): 44-54.
[2] 王庆人, 王银子, 仲红, 张以文. 面向中文的字词组合序列实体识别方法[J]. 清华大学学报(自然科学版), 2023, 63(9): 1326-1338.
[3] 朱唯一, 张雪芹, 顾春华. 基于EDLATrust算法的社交网络信息泄露节点概率预测[J]. 清华大学学报(自然科学版), 2022, 62(2): 355-366.
[4] 王绍卿, 李翠平, 王征, 陈红. 基于多重信任关系的微博转发行为预测[J]. 清华大学学报(自然科学版), 2019, 59(4): 270-275.
[5] 屠守中, 杨婧, 赵林, 朱小燕. 半监督的微博话题噪声过滤方法[J]. 清华大学学报(自然科学版), 2019, 59(3): 178-185.
[6] 芦效峰, 张胜飞, 伊胜伟. 基于CNN和RNN的自由文本击键模式持续身份认证[J]. 清华大学学报(自然科学版), 2018, 58(12): 1072-1078.
[7] 严素蓉, 冯小青, 廖一星. 基于矩阵分解的社会化推荐模型[J]. 清华大学学报(自然科学版), 2016, 56(7): 793-800.
[8] 朱涵钰, 吴联仁, 吕廷杰. 社交网络用户隐私量化研究: 建模与实证分析[J]. 清华大学学报(自然科学版), 2014, 54(3): 402-406.
[9] 韩心慧, 肖祥全, 张建宇, 刘丙双, 张缘. 基于社交关系的DHT网络Sybil攻击防御[J]. 清华大学学报(自然科学版), 2014, 54(1): 1-7.
Full text



版权所有 © 《清华大学学报(自然科学版)》编辑部
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持