基于时空图注意力网络的多变量时序数据异常检测方法

doi:10.3785/j.issn.1008-973X.2025.10.014

浙江大学学报(工学版)

2025, Vol. 59

Issue (10): 2134-2143 DOI: 10.3785/j.issn.1008-973X.2025.10.014

计算机技术

基于时空图注意力网络的多变量时序数据异常检测方法

肖刚(

),卢大鹏,郑文博,程振波,张元鸣*(

)

浙江工业大学计算机科学与技术学院，浙江杭州 310023

Multivariable time series data anomaly detection method based on spatiotemporal graph attention network

Gang XIAO(

),Dapeng LU,Wenbo ZHENG,Zhenbo CHENG,Yuanming ZHANG*(

)

College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou 310023, China

全文: PDF(1128 KB) HTML

摘要：

现有时序数据异常检测方法侧重于提取时序数据的时间变化特征，忽略了多变量之间的空间依赖特征，为此提出基于时空图注意力网络的检测方法. 将多变量时序数据转换为具有时空依赖关系的时序图结构数据，设计时空图注意力网络分别提取时序图结构数据的时间变化特征和空间依赖特征. 通过多层感知机学习时空特征的周期性模式，根据时序数据的预测值与观测值的异常分数进行异常检测. 在公开数据集上的实验结果表明，所提方法在异常检测精准度和鲁棒性方面显著优于现有先进的基线方法.

关键词： 异常检测; 时序图; 图深度学习; 多变量时序数据; 时空特征融合

Abstract:

Existing anomaly detection methods of time series data focus on extracting the temporal variation features, while the spatial dependency features between multiple variables are ignored. To address this problem, a detection method based on a spatiotemporal graph attention network was proposed. The original multivariate time series data were transformed into a time-series graph with spatiotemporal dependencies, and a spatiotemporal graph attention network was designed to separately extract the temporal variation features and spatial dependency features. The periodic patterns of fused spatiotemporal features were learned by a multilayer perceptron, and an anomaly detection was performed based on the anomaly scores between prediction values and observation values. Experimental results on public datasets showed that the proposed method significantly outperformed state-of-the-art baseline methods in terms of anomaly detection accuracy and robustness.

Key words: anomaly detection spatiotemporal graph graph deep learning multivariate time series data spatiotemporal feature fusion

收稿日期: 2024-10-08 出版日期: 2025-10-27

CLC:

TN 92

基金资助: 浙江省“尖兵”“领雁”研发攻关项目（2023C01022）；国家自然科学基金资助项目（62476248）.

通讯作者: 张元鸣 E-mail: xg@zjut.edu.cn;zym@zjut.edu.cn

作者简介: 肖刚（1965—），男，教授，从事智能制造、数据治理研究. ocid.org/0000-0002-0806-2156. E-mail：xg@zjut.edu.cn

	服务
	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	作者相关文章
	肖刚
	卢大鹏
	郑文博
	程振波
	张元鸣

引用本文:

肖刚,卢大鹏,郑文博,程振波,张元鸣. 基于时空图注意力网络的多变量时序数据异常检测方法[J]. 浙江大学学报(工学版), 2025, 59(10): 2134-2143.

Gang XIAO,Dapeng LU,Wenbo ZHENG,Zhenbo CHENG,Yuanming ZHANG. Multivariable time series data anomaly detection method based on spatiotemporal graph attention network. Journal of ZheJiang University (Engineering Science), 2025, 59(10): 2134-2143.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2025.10.014 或 https://www.zjujournals.com/eng/CN/Y2025/V59/I10/2134

图 1 多变量时序图建模过程

图 2 时空图注意力网络模型框架

图 3 TimesNet模型的网络结构

表 1 模型性能评价实验用数据集的统计信息

表 2 不同模型在4个数据集中的性能评价指标

图 4 所提模型在不同传感器上的时序数据预测值与观测值

图 5 所提模型在不同传感器上的时序数据预测值与观测值（异常情况）

表 3 所提模型的模块消融实验结果

图 6 序列长度对异常检测性能的影响

图 7 不同时序图对异常检测性能的影响

1	丁小欧, 于晟健, 王沐贤, 等基于相关性分析的工业时序数据异常检测[J]. 软件学报, 2020, 31 (3): 726- 747 DING Xiaoou, YU Shengjian, WANG Muxian, et al Anomaly detection on industrial time series based on correlation analysis[J]. Journal of Software, 2020, 31 (3): 726- 747
2	SIMMROSS-WATTENBERG F, ASENSIO-PEREZ J I, CASASECA-DE-LA-HIGUERA P, et al Anomaly detection in network traffic based on statistical inference and \alpha-stable modeling[J]. IEEE Transactions on Dependable and Secure Computing, 2011, 8 (4): 494- 509 doi: 10.1109/TDSC.2011.14
3	张圣林, 李东闻, 孙永谦, 等面向云数据中心多语法日志通用异常检测机制[J]. 计算机研究与发展, 2020, 57 (4): 778- 790 ZHANG Shenglin, LI Dongwen, SUN Yongqian, et al Unified anomaly detection for syntactically diverse logs in cloud datacenter[J]. Journal of Computer Research and Development, 2020, 57 (4): 778- 790 doi: 10.7544/issn1000-1239.2020.20190875
4	苏卫星, 朱云龙, 刘芳, 等时间序列异常点及突变点的检测算法[J]. 计算机研究与发展, 2014, 51 (4): 781- 788 SU Weixing, ZHU Yunlong, LIU Fang, et al Outliers and change-points detection algorithm for time series[J]. Journal of Computer Research and Development, 2014, 51 (4): 781- 788 doi: 10.7544/issn1000-1239.2014.20120542
5	HODGE V, AUSTIN J A survey of outlier detection methodologies[J]. Artificial Intelligence Review, 2004, 22 (2): 85- 126 doi: 10.1023/B:AIRE.0000045502.10941.a9
6	AGGARWAL C C, YU P S Outlier detection for high dimensional data[J]. ACM SIGMOD Record, 2001, 30 (2): 37- 46 doi: 10.1145/376284.375668
7	LI W, MAHADEVAN V, VASCONCELOS N Anomaly detection and localization in crowded scenes[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36 (1): 18- 32 doi: 10.1109/TPAMI.2013.111
8	MANEVITZ L M, YOUSEF M One-class svms for document classification[J]. Journal of Machine Learning Research, 2002, 2: 139- 154
9	LAPTEV N, AMIZADEH S, FLINT I. Generic and scalable framework for automated time-series anomaly detection [C]// Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Sydney: ACM, 2015: 1939–1947.
10	SAKURADA M, YAIRI T. Anomaly detection using autoencoders with nonlinear dimensionality reduction [C]// Proceedings of the MLSDA 2014 2nd Workshop on Machine Learning for Sensory Data Analysis. Gold Coast: ACM, 2014: 4–11.
11	ZENATI H, ROMAIN M, FOO C S, et al. Adversarially learned anomaly detection [C]// Proceedings of the IEEE International Conference on Data Mining. Singapore: IEEE, 2018: 727–736.
12	BROWN A, TUOR A, HUTCHINSON B, et al. Recurrent neural network attention mechanisms for interpretable system log anomaly detection [C]// Proceedings of the First Workshop on Machine Learning for Computing Systems. Tempe: ACM, 2018: 1–8.
13	GREFF K, SRIVASTAVA R K, KOUTNÍK J, et al LSTM: a search space odyssey[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28 (10): 2222- 2232 doi: 10.1109/TNNLS.2016.2582924
14	ZHOU H, ZHANG S, PENG J, et al Informer: beyond efficient transformer for long sequence time-series forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35 (12): 11106- 11115 doi: 10.1609/aaai.v35i12.17325
15	KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks [EB/OL]. (2017–02–22)[2024–08–23]. https://arxiv.org/pdf/1609.02907.
16	WEI Y, JANG-JACCARD J, XU W, et al LSTM-autoencoder-based anomaly detection for indoor air quality time-series data[J]. IEEE Sensors Journal, 2023, 23 (4): 3787- 3800 doi: 10.1109/JSEN.2022.3230361
17	XU H, PANG G, WANG Y, et al Deep isolation forest for anomaly detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35 (12): 12591- 12604 doi: 10.1109/TKDE.2023.3270293
18	BARRIENTOS-TORRES D, MARTINEZ-RÍOS E A, NAVARRO-TUCH S A, et al Water flow modeling and forecast in a water branch of Mexico City through ARIMA and transfer function models for anomaly detection[J]. Water, 2023, 15 (15): 2792 doi: 10.3390/w15152792
19	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 and Data Mining. Anchorage: ACM, 2019: 2828–2837.
20	霍纬纲, 梁锐, 李永华基于随机Transformer的多维时间序列异常检测模型[J]. 通信学报, 2023, 44 (2): 94- 103 HUO Weigang, LIANG Rui, LI Yonghua Anomaly detection model for multivariate time series based on stochastic Transformer[J]. Journal on Communications, 2023, 44 (2): 94- 103 doi: 10.11959/j.issn.1000-436x.2023042
21	SHEN L, LI Z, KWOK J T. Timeseries anomaly detection using temporal hierarchical one-class network [C]// Proceedings of the 34th International Conference on Neural Information Processing Systems. Vancouver: ACM, 2020: 13016–13026.
22	ZONG B, SONG Q, MIN M R, et al. Deep autoencoding gaussian mixture model for unsupervised anomaly detection [C]// International Conference on Learning Representations. Vancouver: [s.n.], 2018: 1–19.
23	ABDULAAL A, LIU Z, LANCEWICKI T. Practical approach to asynchronous multivariate time series anomaly detection and localization [C]// Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. [S.l.]: ACM, 2021: 2485–2494.
24	ZHANG Z, LI W, DING W, et al STAD-GAN: unsupervised anomaly detection on multivariate time series with self-training generative adversarial networks[J]. ACM Transactions on Knowledge Discovery from Data, 2023, 17 (5): 1- 18
25	ZHOU B, LIU S, HOOI B, et al. BeatGAN: anomalous rhythm detection using adversarially generated time series [C]// Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence. [S.l.]: International Joint Conferences on Artificial Intelligence Organization, 2019: 4433–4439.
26	LIU Y, DING K, LU Q, et al. Towards self-interpretable graph-level anomaly detection [EB/OL]. (2023–10–25)[2024–09–13]. https://arxiv.org/pdf/2310.16520.
27	DENG A, HOOI B. Graph neural network-based anomaly detection in multivariate time series [C]// The 34th AAAI Conference on Artificial Intelligence. [S.l.]: AAAI, 2021: 4027–4035.
28	ZHENG Y, KOH H Y, JIN M, et al Correlation-aware spatial-temporal graph learning for multivariate time-series anomaly detection[J]. IEEE Transactions on Neural Networks and Learning Systems, 2024, 35 (9): 11802- 11816 doi: 10.1109/TNNLS.2023.3325667
29	DING K, SHU K, SHAN X, et al Cross-domain graph anomaly detection[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022, 33 (6): 2406- 2415 doi: 10.1109/TNNLS.2021.3110982
30	ZHAO H, WANG Y, DUAN J, et al. Multivariate time-series anomaly detection via graph attention network [C]// Proceedings of the IEEE International Conference on Data Mining. Sorrento: IEEE, 2020: 841–850.
31	GUO G, WANG H, BELL D, et al. KNN model-based approach in classification [C]// On The Move to Meaningful Internet Systems 2003: CoopIS, DOA, and ODBASE. Berlin: Springer, 2003: 986–996.
32	PARK D, HOSHI Y, KEMP C C A multimodal anomaly detector for robot-assisted feeding using an LSTM-based variational autoencoder[J]. IEEE Robotics and Automation Letters, 2018, 3 (3): 1544- 1551 doi: 10.1109/LRA.2018.2801475
33	WU H, HU T, LIU Y, et al. TimesNet: temporal 2D-variation modeling for general time series analysis [EB/OL]. (2023–04–12)[2024–08–21]. https://arxiv.org/pdf/2210.02186.
34	MATHUR A P, TIPPENHAUER N O. SWaT: a water treatment testbed for research and training on ICS security [C]// Proceedings of the International Workshop on Cyber-physical Systems for Smart Water Networks. Vienna: IEEE, 2016: 31–36.
35	HUNDMAN K, CONSTANTINOU V, LAPORTE C, et al. Detecting spacecraft anomalies using LSTMs and nonparametric dynamic thresholding [C]// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. London: ACM, 2018: 387–395.

[1]	肖剑,武亮亮,何昕泽,胡欣. 基于异常检测的图像特征匹配算法[J]. 浙江大学学报(工学版), 2025, 59(6): 1140-1147.
[2]	蒋沁诚,陶建峰,王洋洋,张宇磊,刘成良. 基于EWT-LSTM的工业机器人关节异常检测[J]. 浙江大学学报(工学版), 2025, 59(5): 982-994.
[3]	孙菲,厉小润,赵辽英,余绍奇. 基于FrFT变换和全变分正则化的异常检测算法[J]. 浙江大学学报(工学版), 2022, 56(7): 1276-1284.
[4]	张鹏,田子都,王浩. 基于改进生成对抗网络的飞参数据异常检测方法[J]. 浙江大学学报(工学版), 2022, 56(10): 1967-1976.
[5]	刘宇,罗森林,曲乐炜,潘丽敏,张笈. 全特征信息均衡建模的内部威胁人物检测[J]. 浙江大学学报(工学版), 2019, 53(4): 777-784.
[6]	魏媛,冯天恒,黄平捷,侯迪波,张光新. 管网水质多指标动态关联异常检测方法[J]. 浙江大学学报(工学版), 2016, 50(7): 1402-1409.
[7]	何慧梅, 侯迪波, 赵海峰, 黄平捷, 张光新. 基于多因子融合的水质异常检测算法[J]. J4, 2013, 47(4): 735-740.
[8]	黄金钟朱淼良郭晔. 基于文法的异常检测[J]. J4, 2006, 40(2): 243-248.

Viewed

Full text

Abstract

Cited

Shared

Discussed