Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 108 of the 108 references it cites were found in a registry or an index.

A Survey of Deep Anomaly Detection in Multivariate Time Series: Taxonomy, Applications, and Directions

Open access

Abstract

Multivariate time series anomaly detection (MTSAD) can effectively identify and analyze anomalous behavior in complex systems, which is particularly important in fields such as financial monitoring, industrial equipment fault detection, and cybersecurity. MTSAD requires simultaneously analyze temporal dependencies and inter-variable relationships have prompted researchers to develop specialized deep learning models to detect anomalous patterns. In this paper, we conducted a structured and comprehensive overview of the latest techniques in deep learning for multivariate time series anomaly detection methods. Firstly, we proposed a taxonomy for the anomaly detection strategies from the perspectives of learning paradigms and deep learning models, and then provide a systematic review that emphasizes their advantages and drawbacks. We also organized the public datasets for time series anomaly detection along with their respective application domains. Finally, open issues for future research on MTSAD were identified.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.3390/s25010190

Cites

108 of 108 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 4 The general pipeline for MTSAD using deep learning models. Given a source data, we first process the source data using a data processing module that performs basic data cleaning and normalization tasks. Subsequently, we utilize the anomaly detection strategies from the perspectives of learning paradigms and deep learning models to obtain representations for executing anomaly detection tasks across different application domains.
Figure 4 The general pipeline for MTSAD using deep learning models. Given a source data, we first process the source data using a data processing module that performs basic data cleaning and normalization tasks. Subsequently, we utilize the anomaly detection strategies from the perspectives of learning paradigms and deep learning models to obtain representations for executing anomaly detection tasks across different application domains.CC BY · PubMed Central
Figure 1 The overall classification framework diagram for multivariate time series anomaly types.
Figure 1 The overall classification framework diagram for multivariate time series anomaly types.CC BY · PubMed Central
Figure 2 (a) Point-wise anomalies, the red dots indicate anomalies, and (b) Patten-wise anomalies, the red areas represent anomalies.
Figure 2 (a) Point-wise anomalies, the red dots indicate anomalies, and (b) Patten-wise anomalies, the red areas represent anomalies.CC BY · PubMed Central
Figure 3 (a) Global intermetric anomalies. (b) The red-highlighted area on the left indicates local intermetric anomalies, while the red-highlighted area on the right indicates temporal-local intermetric anomalies.
Figure 3 (a) Global intermetric anomalies. (b) The red-highlighted area on the left indicates local intermetric anomalies, while the red-highlighted area on the right indicates temporal-local intermetric anomalies.CC BY · PubMed Central
Figure 5 The examples of each type of anomaly criteria: (a) a forecasting loss; (b) a reconstruction loss; and (c) a contrastive loss.
Figure 5 The examples of each type of anomaly criteria: (a) a forecasting loss; (b) a reconstruction loss; and (c) a contrastive loss.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 108 references

Research works printed but not linked: 22

  1. Vaswani, A. (2017, January 4–9). Attention is all you need. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, CA, USA.
  2. Prabhu, H., Valadi, J., and Arjunan, P. (2024). Generative Adversarial Network with Soft–Dynamic Time Warping and Parallel Reconstruction for Energy Time Series Anomaly Detection. arXiv.
  3. Wang, C., Zhuang, Z., Qi, Q., Wang, J., Wang, X., Sun, H., and Liao, J. (2023, January 10–16). Drift doesn’t matter: Dynamic decomposition with diffusion reconstruction for unstable multivariate time series anomaly detection. Proceedings of the 37th International Conference on Neural Information Processing Systems, New Orleans, LA, USA.
  4. Chen, W., Tian, L., Chen, B., Dai, L., Duan, Z., and Zhou, M. (2022, January 17–23). Deep variational graph convolutional recurrent network for multivariate time series anomaly detection. Proceedings of the 39th International Conference on Machine Learning, Baltimore, MD, USA.
  5. Jeong, Y., Yang, E., Ryu, J.H., Park, I., and Kang, M. (2023). Anomalybert: Self–supervised transformer for time series anomaly detection using data degradation scheme. arXiv.
  6. Zong, B., Song, Q., Renqiang Min, M., Cheng, W., Lumezanu, C., Cho, D., and Chen, H. (May, January 30). Deep autoencoding gaussian mixture model for unsupervised anomaly detection. Proceedings of the International Conference on Learning Representations, Vancouver, BC, Canada.
  7. Wu, X., Qiu, X., Li, Z., Wang, Y., Hu, J., Guo, C., and Yang, B. (2024). CATCH: Channel–Aware Multivariate Time Series Anomaly Detection via Frequency Patching. arXiv.
  8. Liu, S., Yao, D., Fang, L., Li, Z., Li, W., Feng, K., and Bi, J. (2024). AnomalyLLM: Few–Shot Anomaly Edge Detection for Dynamic Graphs using Large Language Models. arXiv.
  9. Zhong, Z., Yu, Z., Yang, Y., Wang, W., and Yang, K. (2024). PatchAD: Patch–Based MLP–Mixer for Time Series Anomaly Detection. arXiv.
  10. Pranavan, T., Sim, T., Ambikapathi, A., and Ramasamy, S. (2022). Contrastive predictive coding for anomaly detection in multi–variate time series data. arXiv.
  11. Orhan, A.E., and Pitkow, X. (2017). Skip connections eliminate singularities. arXiv.
  12. Velickovic, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., and Bengio, Y. (2017). Graph attention networks. arXiv.
  13. Doersch, C. (2016). Tutorial on variational autoencoders. arXiv.
  14. Frogner, C., Zhang, C., Mobahi, H., Araya, M., and Poggio, T.A. (2015, January 7–12). Learning with a Wasserstein loss. Proceedings of the 28th International Conference on Neural Information Processing Systems, Montreal, QC, Canada.
  15. Dong, M., Huang, H., and Cao, L. (2024). Can LLMs Serve As Time Series Anomaly Detectors?. arXiv.
  16. Gong, Z., Tang, Y., and Liang, J. (2023). Patchmixer: A patch–mixing architecture for long–term time series forecasting. arXiv.
  17. Moritz, S., Rehbach, F., Chandrasekaran, S., Rebolledo, M., and Bartz–Beielstein, T. (2018, January 15–19). GECCO Industrial Challenge 2018 Dataset: A water quality dataset for the ‘Internet of Things: Online Anomaly Detection for Drinking Water Quality’Competition. Proceedings of the Genetic and Evolutionary Computation Conference, Kyoto, Japan.
  18. Poorebrahimi, 2021, CICIDS 2017, J. Adv. Signal Process., 5, 105
  19. KDD Cup (2024, December 24). The Third International Knowledge Discovery and Data Mining Tools Competition Dataset KDD Cup 1999 Data. Available online: https://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html.
  20. Filonov, P., Lavrentyev, A., and Vorontsov, A. (2016). Multivariate industrial time series with cyber–attack simulation: Fault detection using an lstm–based predictive data model. arXiv.
  21. Muthukrishna, D., and Gupta, R. (2024). A Classifier–Based Approach to Multi–Class Anomaly Detection Applied to Astronomical Time–Series. arXiv.
  22. Gao, S., Koker, T., Queen, O., Hartvigsen, T., Tsiligkaridis, T., and Zitnik, M. (2024). Units: Building a unified time series model. arXiv.

Cited by

See all 119

Metadata

Authors
Wang, F., Jiang, Y., Zhang, R., Wei, A., Xie, J., & Pang, X.
Affiliations
  • South China Normal University (CN)
  • Guangzhou Panyu Polytechnic (CN)
  • Guangdong Songshan Polytechnic (CN)
Corresponding author
  • Xie, Jingming
  • Pang, Xiongwen
Venue
Sensors
Publisher
MDPI AG
Published
2025-01-01
Volume
25
Issue
1
Pages
190
ISSN
1424-8220 (ISSN-L)
OpenAlex
W4405965607
Cited by
119
References
108
Citations by year
2026: 81 · 2025: 38
FWCI
11.36
Type
journal article
Language
en
Open access
Locations
  • Sensors (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
2021 Key Research Platforms and Projects of Colleges and Universities by the Department of Education of Guangdong Province: Research on the Construction of Knowledge Graph Models for Vocational College Architectural Education Supported by Artificial Intelligence (2021ZDZX1112)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
  • Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
  • Physical Sciences › Computer Science › Computer Networks and Communications › Software System Performance and Reliability
Keywords
multivariate time series anomaly detection, deep learning, temporal dependencies, financial monitoring, industrial fault detection, cybersecurity, open problems

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A Survey of Deep Anomaly Detection in Multivariate Time Series: Taxonomy, Applications, and Directions. https://app.cochat.ai/papers/10.3390/s25010190
The paper
Wang, F., Jiang, Y., Zhang, R., Wei, A., Xie, J., & Pang, X. (2025). A Survey of Deep Anomaly Detection in Multivariate Time Series: Taxonomy, Applications, and Directions. Sensors. https://doi.org/10.3390/s25010190

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.