基于时空特征融合与序列重构的时间序列异常检测

doi:10.3778/j.issn.1673-9418.2411060

摘要/Abstract

摘要： 异常检测是时间序列分析的关键组成部分，旨在识别时间序列中的异常事件。针对传统方法在融合时空相关性、捕捉序列常态分布以及时变特性方面的局限性，提出了一种基于时空特征融合与序列重构的时间序列异常检测模型（AnomNet）。该模型包含时空特征融合网络（STF）、时间序列重构网络和异常检测机制（ADM）三个主要部分。时空特征融合网络结合时域卷积网络和图注意力影响网络，分别捕捉时域长期依赖和全局属性关联，实现时空特征的联合建模；时间序列重构网络采用多层编码器-解码器架构，利用时空融合特征和周期信息，学习样本的正常分布，从而放大重构数据与潜在异常之间的差异；异常检测机制通过拟合重构偏差的尾部分布来进行异常检测，一旦异常分数超过预设阈值，该机制将触发对广义帕累托分布参数的更新，为后续检测提供最新标准。在五种公开数据集上的实验结果验证了AnomNet在时间序列异常检测领域达到先进水平。与OmniAnomaly相比，所提模型的平均性能提升了11.89%。

关键词: 深度学习, 神经网络, 异常检测, 时间序列预测

Abstract: Anomaly detection is a critical component of time series analysis for identifying anomalous events. To address the limitations of traditional methods in integrating spatio-temporal correlations, capturing normal data distributions, and handling time-varying characteristics, this paper proposes a time series anomaly detection model based on spatio-temporal feature fusion and sequence reconstruction (AnomNet). This model comprises three main components: spatio-temporal feature fusion network (STF), time series reconstruction network (TSR), and anomaly detection mechanism (ADM). The STF combines temporal convolutional networks and graph attention influence networks to capture temporal long-term dependencies and global attribute associations, thereby achieving joint modeling of spatio-temporal features. The TSR employs a multi-layer encoder-decoder architecture, utilizes spatio-temporal fused features and cyclical information to learn the normal distribution of samples, which amplifies the discrepancies between reconstructed data and potential anomalies. The ADM detects anomalies by fitting the tail distribution of the reconstruction deviations. Once the anomaly score exceeds a predefined threshold, the mechanism triggers an update of the generalized Pareto distribution parameters, providing the latest standards for subsequent detection. The experimental results on five datasets validate that the AnomNet achieves state-of-the-art performance in the field of time series anomaly detection. Compared with OmniAnomaly, the proposed model shows an average performance improvement of 11.89%.

Key words: deep learning, neural networks, anomaly detection, time series prediction

杨彬, 马廷淮, 黄学坚, 王宇博, 王朝明, 赵博文, 于信. 基于时空特征融合与序列重构的时间序列异常检测[J]. 计算机科学与探索, 2025, 19(9): 2384-2398.

YANG Bin, MA Tinghuai, HUANG Xuejian, WANG Yubo, WANG Zhaoming, ZHAO Bowen, YU Xin. Time Series Anomaly Detection Based on Spatio-Temporal Feature Fusion and Sequence Reconstruction[J]. Journal of Frontiers of Computer Science and Technology, 2025, 19(9): 2384-2398.

参考文献

[1] 杨超城, 严宣辉, 陈容均, 等. 融合双重注意力机制的时间序列异常检测模型[J]. 计算机科学与探索, 2024, 18(3): 740-754.
YANG C C, YAN X H, CHEN R J, et al. Time series anomaly detection model with dual attention mechanism[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(3): 740-754.
[2] 汪军. 基于物联网技术的居民安全配电智能化监测研究[J]. 智能物联技术, 2024, 7(6): 149-152.
WANG J. Research on the intelligent monitoring of residents?? safe power distribution based on the Internet of things technology[J]. Technology of IoT & AI, 2024, 7(6): 149-152.
[3] 朱壮壮, 周治平. 高斯混合生成模型检测健康数据异常[J]. 计算机科学与探索, 2022, 16(5): 1128-1135.
ZHU Z Z, ZHOU Z P. Detection of health data based on Gaussian mixture generative model[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(5): 1128-1135.
[4] 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. New York: ACM, 2014: 4-11.
[5] ZHANG Y X, WANG J D, CHEN Y Q, et al. Adaptive memory networks with self-supervised learning for unsupervised anomaly detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(12): 12068-12080.
[6] SHAUKAT K, ALAM T M, LUO S H, et al. A review of time-series anomaly detection techniques: a step to future perspectives[C]//Advances in Information and Communication: Proceedings of the 2021 Future of Information and Communication Conference. Cham: Springer, 2021: 865-877.
[7] ZHONG J J, WANG D, LI C. A nonparametric health index and its statistical threshold for machine condition monitoring[J]. Measurement, 2021, 167: 108290.
[8] GüNG?R E, ?ZMEN A. Distance and density based clustering algorithm using Gaussian kernel[J]. Expert Systems with Applications, 2017, 69: 10-20.
[9] SCHAFFER A L, DOBBINS T A, PEARSON S A. Interrupted time series analysis using autoregressive integrated moving average (ARIMA) models: a guide for evaluating large-scale health interventions[J]. BMC Medical Research Methodology, 2021, 21(1): 58.
[10] 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. New York: ACM, 2018: 387-395.
[11] LU W N, CHENG Y, XIAO C, et al. Unsupervised sequential outlier detection with deep architectures[J]. IEEE Transactions on Image Processing, 2017, 26(9): 4321-4330.
[12] WANG K F, GOU C, DUAN Y J, et al. Generative adversarial networks: introduction and outlook[J]. IEEE/CAA Journal of Automatica Sinica, 2017, 4(4): 588-598.
[13] SU Y, ZHAO Y J, NIU C H, 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. New York: ACM, 2019: 2828-2837.
[14] 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 and Data Mining. New York: ACM, 2020: 3395-3404.
[15] PANG G S, SHEN C H, CAO L B, et al. Deep learning for anomaly detection[J]. ACM Computing Surveys, 2022, 54(2): 1-38.
[16] DENG A L, HOOI B. Graph neural network-based anomaly detection in multivariate time series[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(5): 4027-4035.
[17] HEWAGE P, BEHERA A, TROVATI M, et al. Temporal convolutional neural (TCN) network for an effective weather forecasting using time-series data from the local weather station[J]. Soft Computing, 2020, 24(21): 16453-16482.
[18] WU H, XU J, WANG J, et al. Autoformer: decomposition transformers with auto-correlation for long-term series forecasting[C]//Advances in Neural Information Processing Systems 34, 2021: 22419-22430.
[19] ZHOU T, MA Z, WEN Q, et al. FEDformer: frequency enhanced decomposed transformer for long-term series forecasting[C]//Proceedings of the 39th International Conference on Machine Learning, 2022: 27268-27286.
[20] MA T H, RONG H, HAO Y S, et al. A novel sentiment polarity detection framework for Chinese[J]. IEEE Transactions on Affective Computing, 2022, 13(1): 60-74.
[21] 华哲邦, 李萌, 赵俊峰, 等. 基于时间序列分析的Web Service QoS预测方法[J]. 计算机科学与探索, 2013, 7(3): 218-226.
HUA Z B, LI M, ZHAO J F, et al. Web service QoS prediction method based on time series analysis[J]. Journal of Frontiers of Computer Science and Technology, 2013, 7(3): 218-226.
[22] YANG S B, DENG Z G, LI X F, et al. A novel hybrid model based on STL decomposition and one-dimensional convolutional neural networks with positional encoding for significant wave height forecast[J]. Renewable Energy, 2021, 173: 531-543.
[23] TAO Z L, WEI Y W, WANG X, et al. MGAT: multimodal graph attention network for recommendation[J]. Information Processing & Management, 2020, 57(5): 102277.
[24] AWAIS M, IQBAL M T B, BAE S H. Revisiting internal covariate shift for batch normalization[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(11): 5082-5092.
[25] ZHOU C, PAFFENROTH R C. Anomaly detection with robust deep autoencoders[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2017: 665-674.
[26] SZIGETI M, FERENCI T, KOVáCS L. The use of peak over threshold methods to characterise blood glucose curves[C]//Proceedings of the 2020 IEEE 14th International Symposium on Applied Computational Intelligence and Informatics. Piscataway: IEEE, 2020: 199-204.
[27] ZHOU H H, MA T H, RONG H, et al. MDMN: multi-task and domain adaptation based multi-modal network for early rumor detection[J]. Expert Systems with Applications, 2022, 195: 116517.
[28] PENG K X, MA T H, JIA L, et al. Enhancing collaboration in heterogeneous multiagent systems through communication complementary graph[J]. IEEE Transactions on Cybernetics, 2024, 54(11): 6881-6894.
[29] ABDULAAL A, LIU Z H, 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. New York: ACM, 2021: 2485-2494.
[30] ALTUN K, BARSHAN B, TUN?EL O. Comparative study on classifying human activities with miniature inertial and magnetic sensors[J]. Pattern Recognition, 2010, 43(10): 3605-3620.
[31] MATHUR A P, TIPPENHAUER N O. SWaT: a water treatment testbed for research and training on ICS security[C]//Proceedings of the 2016 International Workshop on Cyber-physical Systems for Smart Water Networks. Piscataway: IEEE, 2016: 31-36.
[32] ANGRYK R A, MARTENS P C, AYDIN B, et al. Multivariate time series dataset for space weather data analytics[J]. Scientific Data, 2020, 7: 227.
[33] XIA S, SUN W, ZOU X, et al. MFAM-AD: an anomaly detection model for multivariate time series using attention mechanism to fuse multi-scale features[J]. PeerJ Computer Science, 2024, 10: e2201.
[34] BREUNIG M M, KRIEGEL H P, NG R T, et al. LOF: identifying density-based local outliers[J]. ACM SIGMOD Record, 2000, 29(2): 93-104.
[35] SCH?LKOPF B, PLATT J C, SHAWE-TAYLOR J, et al. Estimating the support of a high-dimensional distribution[J]. Neural Computation, 2001, 13(7): 1443-1471.
[36] NG A. Sparse autoencoder[J]. CS294A Lecture Notes, 2011, 72: 1-19.
[37] RUFF L, VANDERMEULEN R, GOERNITZ N, et al. Deep one-class classification[C]//Proceedings of the 35th International Conference on Machine Learning, 2018: 4393-4402.
[38] ZONG B, SONG Q, MIN M R, et al. Deep autoencoding Gaussian mixture model for unsupervised anomaly detection[C]//Proceedings of the 6th International Conference on Learning Representations, 2018.
[39] ZHAO H, WANG Y J, DUAN J Y, et al. Multivariate time-series anomaly detection via graph attention network[C]//Proceedings of the 2020 IEEE International Conference on Data Mining. Piscataway: IEEE, 2020: 841-850.
[40] QIU C, PFROMMER T, KLOFT M, et al. Neural transformation learning for deep anomaly detection beyond images[C]//Proceedings of the 38th International Conference on Machine Learning, 2021: 8703-8714.
[41] LI Z, ZHAO Y, HU X Y, et al. ECOD: unsupervised outlier detection using empirical cumulative distribution functions[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(12): 12181-12193.
[42] XU H Z, WANG Y J, JIAN S L, et al. Calibrated one-class classification for unsupervised time series anomaly detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2024, 36(11): 5723-5736.
[43] NAM Y, YOON S, SHIN Y, et al. Breaking the time-frequency granularity discrepancy in time-series anomaly detection[C]//Proceedings of the ACM Web Conference 2024. New York: ACM, 2024: 4204-4215.
[44] LAI C Y, SUN F K, GAO Z Q, et al. Nominality score conditioned time series anomaly detection by point/sequential reconstruction[C]//Advances in Neural Information Processing Systems 36, 2024.
[45] ZHANG Z W, WANG R Q, DING R, et al. Unravel anomalies: an end-to-end seasonal-trend decomposition approach for time series anomaly detection[C]//Proceedings of the 2024 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway: IEEE, 2024: 5415-5419.
[46] LAI K H, ZHA D, XU J, et al. Revisiting time series outlier detection: definitions and benchmarks[C]//Proceedings of the 35th Conference on Neural Information Processing Systems Datasets and Benchmarks Track (round 1), 2021.