计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (8): 45-51.doi: 10.11896/jsjkx.220600160

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于张量加权与截断核范数的交通数据修复方法

武江南, 张红梅, 赵永梅, 曾航, 胡钢   

  1. 空军工程大学装备管理与无人机工程学院 西安 710051
  • 收稿日期:2022-12-27 修回日期:2023-01-30 出版日期:2023-08-15 发布日期:2023-08-02
  • 通讯作者: 张红梅(zhm_plum@163.com)
  • 作者简介:(289258346@qq.com)
  • 基金资助:
    国家自然科学基金(62002381)

Traffic Data Restoration Method Based on Tensor Weighting and Truncated Nuclear Norm

WU Jiangnan, ZHANG Hongmei, ZHAO Yongmei, ZENG Hang, HU Gang   

  1. College of Equipment Management and Unmanned Aerial Vehicle Engineering,Air Force Engineering University,Xi'an 710051,China
  • Received:2022-12-27 Revised:2023-01-30 Online:2023-08-15 Published:2023-08-02
  • About author:WU Jiangnan,born in 1998,postgra-duate.His main research interests include data mining and machine lear-ning.
    ZHANG Hongmei,born in 1970,Ph.D,professor.Her main research interests include information assurance and systems engineering.
  • Supported by:
    National Natural Science Foundation of China(62002381).

PDF (PC)

1831

摘要： 数据缺失问题严重影响了智能交通系统中通过数据监控交通态势、预测交通流量、部署交通规划等一系列活动。为此,运用基于张量奇异值分解的低秩张量补全框架提出了加权与截断核范数相结合的交通流数据重构模型WLRTC-TTNN(Low Rank Tensor Completion of Weighted and Truncated Nuclear Norm),该模型可以有效地对缺失的时空交通数据进行修复。WLRTC-TTNN方法主要有两方面的优点:一是加入权重因子解决了原始模型对数据输入方向的依赖问题,实现了模型方向的灵活性;二是运用张量的截断核范数来代替张量的核范数作为张量秩最小化的凸代理,保留了时空交通数据内部主要的特征信息,且根据广义奇异值阈值理论,对较小奇异值进行惩罚处理,进一步优化了模型,最终使用交替乘子法实现了WLRTC-TTNN算法。在两个公开的时空交通数据集上选取不同的缺失场景与缺失率进行实验,结果表明:WLRTC-TTNN的补全性能优于其他基线模型,整体的补全精度提高了3%~37%,在数据极端缺失的情况下,其补全效果更加稳定。

关键词: 智能交通, 数据修复, 张量加权, 截断核范数, 交替乘子法

Abstract: The problem of missing data seriously affects a series of activities in intelligent transportation systems,such as monitoring traffic dynamics,predicting traffic flow,and deploying traffic planning through data.Therefore,a traffic flow data reconstruction model WLRTC-TTNN(low rank tensor completion of weighted and truncated nuclear norm)combined with weighted and truncated nuclear norm is proposed by using the low-rank tensor completion framework based on tensor singular value decomposition,which can effectively repair the missing spatio-temporal traffic data.The truncated nuclear norm of the tensor is used as a convex proxy for tensor rank minimization instead of tensor rank minimization,which preserves the main feature information inside the spatio-temporal traffic data,and further optimizes the model by penalizing smaller singular values according to the gene-ralized singular value threshold theory,and finally the WLRTC-TTNN algorithm is implemented using the alternating multiplier method.Experiments are conducted on two publicly available spatio-temporal traffic datasets selected with different missing scenarios and missing rates,and the results show that the complementary performance of WLRTC-TTNNN is better than that of other baseline models,and the overall complementary accuracy improves by 3%~37%,and the complementary effect is more stable in extreme missing scenarios.

Key words: Intelligent transportation, Data repair, Tensor weighting, Truncated kernel norm, Alternating multiplier method

中图分类号: 

  • TP274.2
[1]LI H,LI M,LIN X,et al.A spatiotemporal approach for traffic data imputation with complicated missing patterns[J].Transportation Research Part C Emerging Technologies,2020,119(11):102730.1-102730.23.
[2]LI Q,TAN H,WU Y,et al.Traffic Flow Prediction With Mis-sing Data Imputed by Tensor Completion Methods[J].IEEE Access,2020,8:63188-63201.
[3]YUAN H,CHEN Z H.Short-term Traffic Flow PredictionBased on Temporal Convolutional Networks[J].Journal of South China University of Technology(Natural Science Edition),2020,48(11):107-113.
[4]YU H F,RAO N,DHILLON I S.Temporal regularized matrix factorization for high-dimensional time series prediction[C]//NeurIPS.2016:847-855.
[5]CHEN X B,CHEN C,CHEN L,et al.Missing value interpolation of traffic volume data based on improved low-rank matrix completion [J].Journal of Traffic and Transportation Enginee-ring,2019,19(5):180-190.
[6]SUN L,CHEN X.Bayesian temporal factorization for multidi-mensional time series prediction[J].arXiv:1910.06366,2021.
[7]CHEN X B,LIANG S R,KE J,et al.Traffic Data Recovery Method based on Graph regularization and Schatten-p norm [J].Journal of Southwest Jiaotong University,2022(6):1326-1333.
[8]TAN H,FENG G,FENG J,et al.A tensor-based method for missing traffic data completion[J].Transportation Research Part C,2013,28:15-27.
[9]LIU J,MUSIALSK P,WONKA P,et al.Tensor completion for estimating missing values in visual data[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(1):208-220.
[10]CHEN X,YANG J,SEN L.A Nonconvex Low-Rank TensorCompletion Model for Spatiotemporal Traffic Data Imputation[J].Transportation Research Part C:Emerging Technologies,2020,117(8),102673.1-102673.12.
[11]CHEN X Y,CHEN Y X,SAUNIER N,et al.Scalable low-rank tensor learning for spatiotemporal traffic data imputation[J].Transportation Research Part C:Emerging Technologies,2021,129,103226.1-103226.13.
[12]CHEN X,LEI M,SAUNIER N,et al.Low-Rank Autoregressive Tensor Completion for Spatiotemporal Traffic Data Imputation[J].arXiv:2104.14936,2021.
[13]ZHANG Z M,AERON S.Exact Tensor Completion Using t-SVD[J].IEEE Transactions on Signal Processing,2017(6):1511-1526.
[14]LIU C S,SHAN H,CHEN C L.Tensor p-shrinkage nuclearnorm for low-rank tensor completion[J].Neurocomputing,2020,387:255-267.
[15]LU C,FENG J,CHEN Y,et al.Tensor Robust Principal Component Analysis with A New Tensor Nuclear Norm[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,42(4):925-938.
[16]SONG Y,LI J,CHEN X,et al.An efficient tensor completion method via truncated nuclear norm[J].Journal of Visual Communication and Image Representation,2020,70:102791.1-102791.8.
[17]SU Y,WU X,LIU W.Low-rank Tensor Completion by Sum of Tensor Nuclear Norm Minimization[J].IEEE Access,2019,7:134943-134953.
[18]WANG Y,YIN W,ZENG J.Global Convergence of ADMM in Nonconvex Nonsmooth Optimization[J].arXiv:1511.06324,2018.
[19]BENTBI A H,HACHIMI A E,JBILOU K,et al.On the tensor nuclear norm and the total variation regularization for image and video completion[J].arXiv:2102.10393,2021.
[20]ZUO W,MENG D,ZHANG L,et al.A generalized iteratedshrinkage algorithm for non-convex sparse coding[C]//2013 IEEE International Conference on Computer Vision.2013:1-8.
[21]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.
[22]TAN H C,YANG Z X,FENG G D,et al.Correlation Analysis for Tensor-based Traffic Data Imputation Method[J].Procedia-Social and Behavioral Sciences,2013,96:2611-2620.
[23]BATTAGLINO C,BALLARD G,KOLDA T G.A practical randomized CP tensor decomposition [J].SIAM Journal on Matrix Analysis and Applications,2018,39:876-901.
[24]CHEN X,HE Z,SUN L.A Bayesian tensor decomposition approach for spatiotemporal traffic data imputation[J].Transportation Research Part C:Emerging Technologies,2019,98:73-84.
[25]CHEN X,HE Z,WANG J,et al.Spatial-temporal traffic speed patterns discovery and incomplete data recovery via svd-combined tensor decomposition[J].Transportation Research Part C:Emerging Technologies,2018,86:59-77.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路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