计算机科学 ›› 2023, Vol. 50 ›› Issue (4): 40-46.doi: 10.11896/jsjkx.220200079

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于多邻接图与多头注意力机制的短期交通流量预测

尹恒1, 张凡1,2, 李天瑞1,2,3   

  1. 1 西南交通大学计算机与人工智能学院 成都 611756
    2 四川省制造业产业链协同与信息化支撑技术重点实验室 成都 611756
    3 综合交通大数据应用技术国家工程实验室 成都 611756
  • 收稿日期:2022-02-16 修回日期:2022-09-05 出版日期:2023-04-15 发布日期:2023-04-06
  • 通讯作者: 张凡(fan.zhang@swjtu.edu.cn)
  • 作者简介:(yinheng714995@163.com)
  • 基金资助:
    国家自然科学基金(61773324);四川省重点研发项目(2020YFG0035)

Short-time Traffic Flow Forecasting Based on Multi-adjacent Graph and Multi-head Attention Mechanism

YIN Heng1, ZHANG Fan1,2, LI Tianrui1,2,3   

  1. 1 School of Computing and Artificial Intelligence,Southwest Jiaotong University,Chengdu 611756,China
    2 Manufacturing Industry Chains Collaboration and Information Support Technology Key Laboratory of Sichuan Province,Chengdu 611756,China
    3 National Engineering Laboratory of Integrated Transportation Big Data Application Technology,Chengdu 611756,China
  • Received:2022-02-16 Revised:2022-09-05 Online:2023-04-15 Published:2023-04-06
  • About author:YIN Heng,born in 1997,postgraduate.His main research interests include traffic flow prediction and urban computing.
    ZHANG Fan,born in 1989,Ph.D,research assistant.His main research interests include fault diagnosis,pattern recognition and data fusion.
  • Supported by:
    National Natural Science Foundation of China(61773324) and Sichuan Key R&D Project,China(2020YFG0035).

摘要: 交通流预测在智慧城市系统中占有重要地位,是许多交通方向应用的基石。该任务的难点在于如何有效地建模交通流的时空依赖。现有方法大都使用图卷积网络(Graph Convolution Networks,GCN)建模空间关系,使用卷积神经网络网络(Convolution Neural Network,CNN)或者循环神经网络(Recurrent Neural Network,RNN)建模时间关系,但在建模空间关系时往往只利用邻接矩阵建模了局部关系而忽略了全局空间信息。而在整个路网中存在一些道路,其周围的路网结构相似,这些道路在路网中承载的作用是相似的,这些相似道路的特征也可以作为流量预测的依据。因此,提出一种基于多邻接图与多头注意力机制的时空网络模型MA-STGCN,包括:1)利用node2vec算法计算路网中道路的向量表示,通过阈值计算出相似矩阵用于图卷积操作,抽取全局空间信息;2)利用多通道自注意力机制深入挖掘模型的时空特征。在公开数据集PEMS04与PEMS08上进行的实验验证了该模型的有效性,其准确率与主流的模型相比均有提高。

关键词: 交通流预测, 空洞卷积, 时空网络, 注意力机制, 节点嵌入

Abstract: Traffic flow forecasting is the cornerstone of many applications in transportation which has a great importance in smart city system.The difficulty of this task is how to effectively model the temporal and spatial dependence.Existing methods usually use GNN to model temporal correlation and CNN or RNN to model temporal correlation.When modeling the spatial correlation,only the adjacency matrix is applied to model local relationships while ignoring global spatial information.However,there are some roads in the entire road network whose surrounding structures are similar,and these roads carry similar functions in the road network.Therefore,the characteristics of these similar roads can also be used as the basis for traffic prediction.This paper proposes a traffic flow forecasting model based on multi-adjacent matrix and multi-head attention mechanism.It includes:1)the node2vec algorithm is applied to calculate the vector representation of the road in road network,and the similarity matrix is calculated through the threshold for graph convolution operation to extract global spatial information;2)the multi-channel self-attention mechanism is used to mine the spatial and temporal features of the model.Experiments on public datasets PEMS04 and PEMS08 demonstrate the proposed model’s effectiveness.Its accuracy is improved compared with the mainstream models.

Key words: Traffic forecasting, Dilated convolution, Spatial-Temporal network, Attention mechanism, Node embedding

中图分类号: 

  • TP391
[1]AGARAP A F M.A neural network architecture combining gated recurrent unit(GRU)and support vector machine(SVM)for intrusion detection in network traffic data[C]//Proceedings of the 2018 10th International Conference on Machine Learning and Computing.2018:26-30.
[2]WILLIAMS B M,HOEL L A.Modeling and forecasting vehicular traffic flow as a seasonal arima process:Theoretical basis and empirical results[J].Journal of Transportation Engineering,2003,129(6):664-672.
[3]LESHEM G,RITOV Y.Traffic flow prediction using adaboost algorithm with random forests as a weak learner[C]//Procee-dings of World Academy of Science,Engineering and Technology.2007:193-198.
[4]LECUN Y,BENGIO Y,HINTON G.Deep learning[J].Nature,2015,521(7553):436-444.
[5]CUI Z,KE R,PU Z,et al.Deep bidirectional and unidirectional LSTM recurrent neural network for network-wide traffic speed prediction[J].arXiv:1801.02143,2018.
[6]ZHANG X H,FENG A M.Short-Time traffic flow prediction based on and LSTM network[J].Application Research of Computers,2021,41(1):225-230.
[7]SHI X J,CHEN Z,WANG H,et al.Convolutional LSTM network:A machine learning approach for precipitation nowcasting[C]//Advances inNeural Information Processing Systems.2015:802-810.
[8]ZHANG J,ZHENG Y,QI D.Deep spatio-temporal residual networks for citywide crowd flows prediction[C]//Proceedings of Thirty-first AAAI Conference on Artificial Intelligence.2017:1566-1661.
[9]ESTRACH J B,ZAREMBA W,SZLAM A,et al.Spectral networks and deep locally connected networks on graphs[C]//Proceedings of 2nd International Conference on Learning Representations.2014:1-14.
[10]SEO Y,DEFFERRARD M,VANDERGHEYNST P,et al.Struc-tured sequence modeling with graph convolutional recurrent networks[C]//Proceedings of International Conference on Neural Information Processing.Cham:Springer,2018:362-373.
[11]LI Y,YU R,SHAHABI C,et al.Diffusion convolutional recurrent neural network:Data-driven traffic forecasting[C]//Proceedings of International Conference on Learning Representation.2018:1-16.
[12]SUTSKEVER I,VINYALS O,LE Q V.Sequence to sequencelearning with neural networks[C]//Advances inNeural Information Processing Systems.2014:3104-3112.
[13]XIONG T,QI Y,ZHANG W B.Short-Time traffic flow prediction in road network based on DCGRU-RF model[J].Computer Science,2020,47(5):84-89.
[14]YU B,YIN H,ZHU Z.Spatio-temporal graph convolutional networks:A deep learning framework for traffic forecasting[J].arXiv:1709.04875,2017.
[15]GUO S,LIN Y,FENG N,et al.Attention based spatial-temporal graph convolutional networks for traffic flow forecasting[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2019:922-929.
[16]WU Z,PAN S,LONG G,et al.Graph wavenet for deep spatial-temporal graph modeling[J].arXiv:1906.00121,2019.
[17]YU F,KOLTUN V.Multi-scale context aggregation by dilated convolutions[J].arXiv:1511.07122,2015.
[18]LI H,WANG F,XIE S Y,et al.A traffic prediction model with dual autoregressive components based on improved graph wave network[J].Computer Science,2021,48(S2):159-165,183.
[19]SONG C,LIN Y,GUO S,et al.Spatial-temporal synchronousgraph convolutional networks:A new framework for spatial-temporal network data forecasting[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2020:914-921.
[20]PEROZZI B,AL-RFOU R,SKIENA S.Deepwalk:Online lear-ning of social representations[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Disco-very and Data Mining.2014:701-710.
[21]MIKOLOV T,SUTSKEVER I,CHEN K,et al.Distributed representations of words and phrases and their compositionality[C]//Advances in Neural Information Processing Systems.2013:3111-3119.
[22]TANG J,QU M,WANG M,et al.Line:Large-scale information network embedding[C]//Proceedings of the 24th International Conference on World Wide Web.2015:1067-1077.
[23]GROVER A,LESKOVEC J.node2vec:Scalable feature learning for networks[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mi-ning.2016:855-864.
[24]VASWANI A,SHAZEER N,PARMAR N,et al.Attention isall you need[C]//Advances inNeural Information Processing Systems.2017:5998-6008.
[25]WEISS M,BONNEL P,HUMMEL R,et al.Analyzing on-road emissions of light-duty vehicles with Portable Emission Mea-surement Systems(PEMS)[R].JRC Scientific and Technical Reports,EUR,2011.
[26]DEFFERRARD M,BRESSON X,VANDERGHEYNST P.Con-volutional neural networks on graphs with fast localized spectral filtering[J].Advances in Neural Information Processing Systems,2016,29:3844-3852.
[27]SRIVASTAVA N,HINTON G,KRIZHEVSKY A,et al.Dropout:a simple way to prevent neural networks from overfitting[J].Journal of Machine Learning Research,2014,15(1):1929-1958.
[1] 韩雪明, 贾彩燕, 李轩涯, 张鹏飞.
传播树结构结点及路径双注意力谣言检测模型
Dual-attention Network Model on Propagation Tree Structures for Rumor Detection
计算机科学, 2023, 50(4): 22-31. https://doi.org/10.11896/jsjkx.220200037
[2] 雒晓辉, 吴云, 王晨星, 余文婷.
基于用户长短期偏好的序列推荐模型
Sequential Recommendation Model Based on User’s Long and Short Term Preference
计算机科学, 2023, 50(4): 47-55. https://doi.org/10.11896/jsjkx.220100264
[3] 王娅丽, 张凡, 余增, 李天瑞.
基于交互注意力和图卷积网络的方面级情感分析
Aspect-level Sentiment Classification Based on Interactive Attention and Graph Convolutional Network
计算机科学, 2023, 50(4): 196-203. https://doi.org/10.11896/jsjkx.220100105
[4] 于兴崭, 芦天亮, 杜彦辉, 王曦锐, 杨成.
基于合成图像和Xception改进模型的安卓恶意家族分类方法
Android Malware Family Classification Method Based on Synthetic Image and Xception Improved Model
计算机科学, 2023, 50(4): 351-358. https://doi.org/10.11896/jsjkx.220300200
[5] 李帅, 徐彬, 韩祎珂, 廖同鑫.
SS-GCN:情感增强和句法增强的方面级情感分析模型
SS-GCN:Aspect-based Sentiment Analysis Model with Affective Enhancement and Syntactic Enhancement
计算机科学, 2023, 50(3): 3-11. https://doi.org/10.11896/jsjkx.220700238
[6] 陈富强, 寇嘉敏, 苏利敏, 李克.
基于图神经网络的多信息优化实体对齐模型
Multi-information Optimized Entity Alignment Model Based on Graph Neural Network
计算机科学, 2023, 50(3): 34-41. https://doi.org/10.11896/jsjkx.220700242
[7] 周明强, 代开浪, 吴全旺, 朱庆生.
异构信息网络的注意力感知多通道图卷积评分预测模型
Attention-aware Multi-channel Graph Convolutional Rating Prediction Model for Heterogeneous Information Networks
计算机科学, 2023, 50(3): 129-138. https://doi.org/10.11896/jsjkx.220300004
[8] 冯程程, 刘派, 姜琳颖, 梅笑寒, 郭贵冰.
文档增强型知识库问答
Document-enhanced Question Answering over Knowledge-Bases
计算机科学, 2023, 50(3): 266-275. https://doi.org/10.11896/jsjkx.220300022
[9] 邹芸竹, 杜圣东, 滕飞, 李天瑞.
一种基于多模态深度特征融合的视觉问答模型
Visual Question Answering Model Based on Multi-modal Deep Feature Fusion
计算机科学, 2023, 50(2): 123-129. https://doi.org/10.11896/jsjkx.211200303
[10] 王鹏宇, 台文鑫, 刘芳, 钟婷, 罗绪成, 周帆.
基于数据增强的自监督飞行航迹预测
Self-supervised Flight Trajectory Prediction Based on Data Augmentation
计算机科学, 2023, 50(2): 130-137. https://doi.org/10.11896/jsjkx.211200016
[11] 瞿中, 王彩云.
基于注意力机制和轻量级空洞卷积的混凝土路面裂缝检测
Crack Detection of Concrete Pavement Based on Attention Mechanism and Lightweight DilatedConvolution
计算机科学, 2023, 50(2): 231-236. https://doi.org/10.11896/jsjkx.211200290
[12] 刘露平, 周欣, 程军军, 何小海, 卿粼波, 王美玲.
基于会话式机器阅读理解模型的事件抽取方法
Event Extraction Method Based on Conversational Machine Reading Comprehension Model
计算机科学, 2023, 50(2): 275-284. https://doi.org/10.11896/jsjkx.220400271
[13] 蔡肖, 陈志华, 盛斌.
基于移位窗口金字塔Transformer的遥感图像目标检测
SPT:Swin Pyramid Transformer for Object Detection of Remote Sensing
计算机科学, 2023, 50(1): 105-113. https://doi.org/10.11896/jsjkx.211100208
[14] 张婧媛, 王宏霞, 何沛松.
基于Transformer的多任务图像拼接篡改检测算法
Multitask Transformer-based Network for Image Splicing Manipulation Detection
计算机科学, 2023, 50(1): 114-122. https://doi.org/10.11896/jsjkx.211100269
[15] 李雪辉, 张拥军, 史殿习, 徐化池, 史燕燕.
融合注意力特征的无锚框视觉目标跟踪方法
AFTM:Anchor-free Object Tracking Method with Attention Features
计算机科学, 2023, 50(1): 138-146. https://doi.org/10.11896/jsjkx.211000083
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!