Search

Multivariate Time Series Anomaly Detection Based on Time-Domain and Frequency-Domain Fusion

Author(s)

Jiming Xu, Hanqi Liu, Jun Xu

Affiliation(s)

College of Science, North China University of Science and Technology, Tangshan, China

Abstract

In complex multivariate time-series anomaly detection tasks, traditional methods often rely on time-domain modeling while neglecting frequency-domain information, leading to limited performance. To address this issue, this study proposes an innovative time-frequency anomaly detection framework, FTAD, which combines time-domain and frequency-domain features through adaptive graph attention and frequency-domain attention mechanisms, significantly improving anomaly detection performance for multivariate time-series data. the model first decomposes the time-series data using Exponential Moving Average (EMA), then processes the time-domain and frequency-domain features separately using Adaptive Graph Attention (AGAT) and Frequency-domain Attention (FreRA), respectively. Finally, the mutual inverse entropy weighting (REFusion) mechanism is used to dynamically and adaptively fuse time-domain and frequency-domain features. Experimental results show that FTAD outperforms existing methods on multiple datasets, demonstrating its effectiveness in complex time-series data.

Keywords

Industrial Internet of Things; Multivariate Time Series Data; Time-Domain Feature; Frequency-Domain Feature; Adaptively Fusion

References

[1] Wen Q, Yang L, Zhou T, et al. Robust time series analysis and applications: An industrial perspective [C] // Proc of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2022:4836-4837. [2] Kheddar H, Himeur Y, Awad A I. Deep transfer learning for intrusion detection in industrial control networks: A comprehensive review [J]. Journal of Network and Computer Applications, 2023, 220:103760. [3] Wang F, Jiang Y, Zhang R, et al. A Survey of Deep Anomaly Detection in Multivariate Time Series: Taxonomy, Applications, and Directions [J]. Sensors (Basel, Switzerland), 2025, 25(1):190. [4] Si H, Li J, Pei C, et al. Timeseriesbench: An industrial-grade benchmark for time series anomaly detection models [C]//2024 IEEE 35th International Symposium on Software Reliability Engineering (ISSRE). IEEE, 2024:61-72. [5] Wang Z, Pei C, Ma M, et al. Revisiting VAE for unsupervised time series anomaly detection: A frequency perspective [C] // Proceedings. [6] Zhou T, Ma Z, Wen Q, et al. Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting [C] // International Conference on Machine Learning. PMLR, 2022:27268-27286. [7] Lu Y X, Jin X B, Chen J, et al. F-SE-LSTM: A Time Series Anomaly Detection Method with Frequency Domain Information [J/OL]. arXiv:2412.02474, 2024[2025-08-11]. https://arxiv. org/abs/2412.02474 [8] Su Y, Zhao Y, Niu C, et al. Robust anomaly detection for multivariate time series through stochastic recurrent neural network [C] // Proc of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM, 2019:2828-2837. [9] Hundman K, Constantinou V, Laporte C, et al. Detecting spacecraft anomalies using LSTMs and nonparametric dynamic thresholding [C] // Proc of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM, 2018:387-395. [10]. 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. [11]. SU Y, ZHAO Y J, NIU CH H, et al. Robust anomalydetectionformultivariatetimeseriesthrough stochasticrecurrentneuralnetwork [C]. 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2019:2828-2837. [12]. Audibert J, Michiardi P, Guyard F, et al. USAD: Unsupervised anomaly detection on multivariate time series [C] // Proc of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM, 2020:3395-3404. [13]. Tuli S, Casale G, Jennings N R. TranAD: Deep transformer networks for anomaly detection in multivariate time series data [J/OL]. arXiv:2201.07284, 2022[2025-08-11]. https://arxiv. org/abs/2201.07284. [14]. Deng A, Hooi B. Graph neural network-based anomaly detection in multivariate time series [C] // Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto, CA: AAAI Press, 2021, 35(5):4027-4035. [15]. Fang Y, Xie J, Zhao Y, et al. Temporal-frequency masked autoencoders for time series anomaly detection [C]//2024 IEEE 40th International Conference on Data Engineering (ICDE). IEEE, 2024:1228-1241. [16]. ZHAO H, WANG Y J, DUAN J Y, et al. Multivariatetime-seriesanomaly detection via graph attention network [C]. 2020 IEEE International ConferenceonData Mining(ICDM), 2020:841-850.