计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (6A): 240700124-8.doi: 10.11896/jsjkx.240700124

• 大数据&数据科学 • 上一篇    下一篇

基于变分图自编码器的多变量时序数据异常检测

尹文萃, 谢平, 叶成绪, 韩佳新, 夏星   

  1. 青海师范大学计算机学院 西宁 810016
    青海师范大学青海省物联网重点实验室 西宁 810008
    青海师范大学省部共建藏语信息智能处理及应用国家重点实验室 西宁 810008
  • 出版日期:2025-06-16 发布日期:2025-06-12
  • 通讯作者: 谢平(eping@qhnu.edu.cn)
  • 作者简介:(1642333065@qq.com)
  • 基金资助:
    国家自然科学基金项目(62362057);西宁市科学技术局项目(2024-Y-8)

Anomaly Detection of Multi-variable Time Series Data Based on Variational Graph Auto-encoders

YIN Wencui, XIE Ping, YE Chengxu, HAN Jiaxin, XIA Xing   

  1. School of Computer,Qinghai Normal University,Xining 810016,China
    Key Laboratory of Internet of Things of Qinghai Province,Qinghai Normal University,Xining 810008,China
    The State Key Laboratory of Tibetan Intelligent Information Processing and Application,Qinghai Normal University,Xining 810008,China
  • Online:2025-06-16 Published:2025-06-12
  • About author:YIN Wencui,born in 2000,postgraduate.Her main research interest includes anomaly detection.
    XIE Ping,born in 1979,postgraduate,Ph.D supervisor.His main research interests include parallel and distributed file systems,network storage systems,fault-tolerant storage systems.
  • Supported by:
    National Natural Science Foundation of China(62362057) and Xining Science and Technology Bureau(2024-Y-8).

PDF (PC)

114

摘要： 多变量时序数据异常检测指识别多变量时序数据中的异常值。为解决多变量时序数据间的复杂性和内部变量间特征依赖的问题,文中提出了一种基于变分图自编码器的多变量时序数据异常检测方法。首先,使用滑动窗口提取变量嵌入特征,并基于特征相似性构建结构关联关系图,然后将该多变量时序数据间的关联关系通过变分图自编码器进行优化,提高多变量时序数据的结构特征表征能力;其次,通过多头注意力机制提升多变量时序数据不同通道间的特征表示,并和多变量时序数据结构信息进行融合;最后,采用极值理论选取阈值并进行无监督异常检测。实验结果表明,所提模型在SWaT,MSL等数据集上F1分数达到了81.43%和99.67%的结果。

关键词: 异常检测, 多变量时序数据, 图结构学习, 变分图自编码器

Abstract: Multivariate time series data anomaly detection refers to identifying outliers in multivariate time series data.In order to solve the problem of complexity between multi-variable time series data and feature dependence between internal variables,this paper proposes an anomaly detection method for multi-variable time series data based on variational graph autoencoders.Firstly,a sliding window is used to extract variable embedding features,and a structural correlation graph is constructed based on feature similarity.Then the correlation between the multi-variable time series data is optimized through a variational graph autoencoder to improve the structural characteristics of the multi-variable time series data.Secondly,the multi-head attention mechanism is used to improve the feature representation between different channels of multi-variable time series data,which is fused with the structural information of multi-variable time series data.Finally,the extreme value theory is used to select the threshold and perform unsupervised anomaly detection.Experimental results show that the F1 scores of this model reaches 81.43% and 99.67% on SWaT,MSL and other datasets,respectively.

Key words: Anomaly detection, Multivariable time series data, Graph structure learning, Variational graph autoencoder

中图分类号: 

  • TP391
[1]NING W Y.Research on multivariate time series anomaly detection methods [D].Nanchang:East China Jiaotong University,2023.
[2]PANG G,SHEN C,CAN L,et al.Deep learning for anomaly de-tection:A review[J].ACM Computing Surveys(CSUR),2021,54(2):1-38.
[3]YANG C C,YAN X H,CHEN R J,et al.Time Series Anomaly Detection Model with Dual Attention Mechanism [J].Journal of Computer Science and Exploration,2024,18(3):740-754.
[4]YIN R G,WEI S,LI H,et al.Introduction of UnsupervisedLearning Methods in Deep Learning [J].Applications of Computer Systems,2016,25(8):1-7.
[5]ZHOU Y,ZHU W H,FANG Q,et al.Survey of Outlier Detection Methods Based on Clustering [J].Computer Engineering and Applications,2019,57(12):37-45.
[6]JI Q,SUN Y F,HU Y L,et al.Review of Clustering With Deep Learning.Journal of Beijing University of Technology,2021,47(8):912-924.
[7]董洪荣,付亚军,张帅,等.基于欧式距离聚类的自适应过采样方法[J].印刷与数字媒体技术研究,2023(5):26-41.
[8]AHMAD H,KASASBEH B,ALDABAYBAH B,et al.Classbalancing framework for credit card fraud detection based on clustering and similarity-based selection(SBS)[J].International Journal of Information Technology,2023,15(1):325-333.
[9]DONG H R,FU Y J,ZHANG S,et al.Review of anomaly detection algorithms for multidimensional time series [J].Print and Digital Media Technology Research,2023(5):26-41.
[10]WANG S H.Research on Anomaly Detection Method of Multivariate Time Series Based on Deep Learning [D].Harbin:Harbin Engineering University,2021.
[11]LAI J,WANG X D,FORWARD,et al.Review on autoencoder and its application [J].Journal of Communications,2021,42(9):218-230.
[12]KIPF,THOMAS N,WELLING.Variational graph auto-en-coders[C]//NIPS Workshop on Bayesian Deep Learning.Barcelona,Spain,2016:1-3.
[13]ZHENG Y J.Research on Time Series Anomaly DetectionBased on Deep Learning[D].Beijing:Beijing University of Civil Engineering and Architecture,2021.
[14]LI D,CHEN D,JIN B,et al.MAD-GAN:Multivariate anomalydetection for time series data with generative adversarial networks[C]//International Conference on Artificial Neural Networks.Cham:Springer International Publishing,2019:703-716.
[15]SU Y,ZHAO Y,NIU C,et al.Robust anomaly detection for multivariate time series through stochastic recurrent neural network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.Anchorage,AK,USA,2019:2828-2837.
[16]AUDIBERT J,MICHIARDI P,GUYARD F,et al.Usad:Unsupervised anomaly detection on multivariate time series[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.New York,United States,2020:3395-3404.
[17]CHEN W,TIAN L,CHEN B,et al.Deep variational graph convolutional recurrent network for multivariate time series anomaly detection[C]//International Conference on Machine Lear-ning.Baltimore,Maryland,USA,2022:3621-3633.
[18]LIU Y,JIN M,PAN S,et al.Graph self-supervised learning:A survey[J].IEEE Transactions on Knowledge and Data Engineering,2022,35(6):5879-5900.
[19]DENG A,HOOI B.Graph neural network-based anomaly detection in multivariate time series[C]//Proceedings of the AAAI Conference on Artificial Intelligence.Vancouver,Canada,2021,35(5):4027-4035.
[20]ZHOU J,CUI G,HU S,et al.Graph neural networks:A review of methods and applications[J].AI open,2020,1:57-81.
[21]VELIČKOVIĆ P,CUCURULL G,CASANOVA A,et al.Graph attention networks[C]//6th International Conference on Lear-ning Representation.Vancouver,BC,Canada,2018:1-12.
[22]XIE Y,XU Z,ZHANG J,et al.Self-supervised learning of graph neural networks:A unified review[J].IEEE transactions on pattern analysis and machine intelligence,2022,45(2):2412-2429.
[23]LI Z ,JIN X L,ZHUANG C Z,et al.Overview on Graph Based Anomaly Detection [J].Journal of Software,2020,32(1):167-193.
[24]YAN S H,CHEN Z D.GAE-based Unsupervised Anomaly Detection of Multivariable Time Series [J].Applications of Computer Systems,2023,32(5):308-315.
[25]ZHANG Y,LU M M,ZHENG Y J,et al.Student Grade Prediction Based on Graph Auto-Encoder Model [J].Computer Engineering and Applications,2019,57(13):251-257.
[26]CHEN X,GÜTTEL S.Fast and exact fixed-radius neighborsearch based on sorting[J].PeerJ Computer Science,2024,10:1-17.
[27]ZHANG S,TONG H,XU J,et al.Graph convolutional net-works:a comprehensive review[J].Computational Social Networks,2019,6(1):1-23.
[28]LI Z,LI C,XU Z,et al.Frequency analysis of precipitation extremes in Heihe River basin based on generalized Pareto distribution[J].Stochastic environmental research and risk assessment,2014,28:1709-1721.
[29]MATHUR A P,TIPPENHAUER N O.SWaT:A water treatment testbed for research and training on ICS security[C]//International Workshop on Cyber-physical Systems for Smart Water Networks(CySWater).Vienna,Austria,2016:31-36.
[30]TULI S,CASALE G,JENNINGS N R.Tranad:Deep trans-former networks for anomaly detection in multivariate time series data[J].arXiv:2201.07284,2022.
[31]CEMGIL T,GHAISAS S,DVIJOTHAM K,et al.The autoencoding variational autoencoder[J].Advances in Neural Information Processing Systems,2020,33:15077-15087.
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