Unsupervised Anomaly Detection for IoT-Based Multivariate Time Series: Existing Solutions, Performance Analysis and Future Directions
Open accessThe DOI matches its Crossref record; 16 of the 16 references it cites were found in a registry or an index.
Abstract
The recent wave of digitalization is characterized by the widespread deployment of sensors in many different environments, e.g., multi-sensor systems represent a critical enabling technology towards full autonomy in industrial scenarios. Sensors usually produce vast amounts of unlabeled data in the form of multivariate time series that may capture normal conditions or anomalies. Multivariate Time Series Anomaly Detection (MTSAD), i.e., the ability to identify normal or irregular operative conditions of a system through the analysis of data from multiple sensors, is crucial in many fields. However, MTSAD is challenging due to the need for simultaneous analysis of temporal (intra-sensor) patterns and spatial (inter-sensor) dependencies. Unfortunately, labeling massive amounts of data is practically impossible in many real-world situations of interest (e.g., the reference ground truth may not be available or the amount of data may exceed labeling capabilities); therefore, robust unsupervised MTSAD is desirable. Recently, advanced techniques in machine learning and signal processing, including deep learning methods, have been developed for unsupervised MTSAD. In this article, we provide an extensive review of the current state of the art with a theoretical background about multivariate time-series anomaly detection. A detailed numerical evaluation of 13 promising algorithms on two publicly available multivariate time-series datasets is presented, with advantages and shortcomings highlighted.
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.
Integrity signals
Registry
Cites
Statistics
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.
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
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805
Chandola, V., Banerjee, A., & Kumar, V. (2009) · ACM Computing Surveys · doi:10.1145/1541880.1541882 · Cited by 11,687 · cited 3×
“Maximum inter-cluster distance and minimum intra-cluster distance are two criteria for effective clustering [27].”
BackgroundMethodA Comprehensive Survey on Graph Neural Networks
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., & Yu, P. S. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.2978386 · Cited by 11,000 · cited 1×
“Several time-series anomaly detection models have been proposed us-ing transformer architectures: a reconstruction-based transformer architecture for early anomaly detection is proposed in [76]; a transformer-based anomaly detection model (TranAD) with an adversarial training procedure is proposed…”
MethodThe graph neural network model
Scarselli, F., Gori, M., Tsoi, A. C., Hagenbuchner, M., & Monfardini, G. (2009) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2008.2005605 · Cited by 9,582 · cited 1×
“Several time-series anomaly detection models have been proposed us-ing transformer architectures: a reconstruction-based transformer architecture for early anomaly detection is proposed in [76]; a transformer-based anomaly detection model (TranAD) with an adversarial training procedure is proposed…”
MethodLearning long-term dependencies with gradient descent is difficult
Bengio, Y., Simard, P., & Frasconi, P. (1994) · IEEE Transactions on Neural Networks · doi:10.1109/72.279181 · Cited by 8,542 · cited 1×
“Afterward, a threshold is set to identify normal and anomaly data.”
BackgroundBackpropagation through time: what it does and how to do it
Werbos, P. J. (1990) · Proceedings of the IEEE · doi:10.1109/5.58337 · Cited by 4,879 · cited 1×
“The second phase involves an anomaly detector, which calculates the anomaly score from 0 to 1 based on the Euclidean distance between the predicted and actual values.”
BackgroundMcClelland, J. L., Rumelhart, D. E., & AU (1987) · doi:10.7551/mitpress/5237.001.0001 · Cited by 3,315
Deep Learning for Anomaly Detection: A Review
Pang, G., Shen, C., Cao, L., & Hengel, A. V. D. (2022) · ACM Computing Surveys · doi:10.1145/3439950 · Cited by 2,837
An Introduction to Variational Autoencoders
Kingma, D. P., & Welling, M. (2019) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000056 · Cited by 2,609 · cited 2×
“Massive volumes of unlabeled data, generated by IoT devices and sensors in the form of multivariate time series, capture either normal or anomalous behavior of monitored systems/environments.”
BackgroundMethod
Research works printed but not linked: 72
- Misra, S., Roy, C., and Mukherjee, A. (2021). Introduction to Industrial Internet of Things and Industry 4.0, CRC Press. · doi:10.1201/9781003020905
- Munir, M., Siddiqui, S.A., Chattha, M.A., Dengel, A., and Ahmed, S. (2019). FuseAD: Unsupervised Anomaly Detection in Streaming Sensors Data by Fusing Statistical and Deep Learning Models. Sensors, 19. · DOI
- Chalapathy, R., and Chawla, S. (2019). Deep Learning for Anomaly Detection: A Survey. arXiv. · DOI · Paper page
- Cook, 2020, Anomaly Detection for IoT Time-Series Data: A Survey, IEEE Internet Things J., 7, 6481, 10.1109/JIOT.2019.2958185 · DOI · Paper page
- Choi, 2021, Deep Learning for Anomaly Detection in Time-Series Data: Review, Analysis and Guidelines, IEEE Access, 9, 120043, 10.1109/ACCESS.2021.3107975 · DOI · Paper page
- Sgueglia, 2022, A systematic literature review of IoT time series anomaly detection solutions, Future Gener. Comput. Syst., 134, 170, 10.1016/j.future.2022.04.005 · DOI
- Garg, 2022, An Evaluation of Anomaly Detection and Diagnosis in Multivariate Time Series, IEEE Trans. Neural Netw. Learn. Syst., 33, 2508, 10.1109/TNNLS.2021.3105827 · DOI · Paper page
- Harivigneshwar, C.J., Dharmavenkatesan, K.B., Ajith, R., and Jeyanthi, R. (2019, January 22–23). Modeling of Multivariate Systems using Vector Autoregression(VAR). Proceedings of the 2019 Innovations in Power and Advanced Computing Technologies, i-PACT 2019, Vellore, India.
- Melnyk, 2016, Vector autoregressive model-based anomaly detection in aviation systems, J. Aerosp. Inf. Syst., 13, 161
- Goyal, V., Xu, M., and Kapat, J. (2019, January 17–21). Use of Vector Autoregressive Model for Anomaly Detection in Utility Gas Turbines. Proceedings of the ASME Turbo Expo, Phoenix, AR, USA. · doi:10.1115/gt2019-90995
- Montgomery, D.C. (2019). Introduction to Statistical Quality Control, John Wiley & Sons Inc.. [8th ed.].
- Lopes, 2004, Comparative study of the performance of the CuSum and EWMA control charts, Comput. Ind. Eng., 46, 707, 10.1016/j.cie.2004.05.025 · DOI
- Lowry, 1995, A review of multivariate control charts, IIE Trans., 27, 800, 10.1080/07408179508936797 · doi:10.1080/07408179508936797
- Mishra, 2021, Detection of data anomalies at the edge of pervasive IoT systems, Computing, 103, 1657, 10.1007/s00607-021-00927-9 · doi:10.1007/s00607-021-00927-9
- Crosier, 1988, Multivariate Generalizations of Cumulative Sum Quality-Control Schemes, Technometrics, 30, 291, 10.1080/00401706.1988.10488402 · DOI · Paper page
- Shyu, M.L., Chen, S.C., Sarinnapakorn, K., and Chang, L. (2003, January 19–22). A Novel Anomaly Detection Scheme Based on Principal Component Classifier. Proceedings of the International Conference on Data Mining, Melbourne, FL, USA.
- Golyandina, N., and Zhigljavsky, A. (2020). Singular Spectrum Analysis for Time Series, Springer. Springer Briefs in Statistics. · DOI · Paper page
- Golyandina, N., Nekrutkin, V., and Zhigljavsky, A.A. (2001). Analysis of Time Series Structure, Chapman and Hall/CRC. · DOI · Paper page
- Hassani, 2013, Multivariate singular spectrum analysis: A general view and new vector forecasting approach, Int. J. Energy Stat., 01, 55, 10.1142/S2335680413500051 · doi:10.1142/s2335680413500051
- Moskvina, 2003, An algorithm based on singular spectrum analysis for change-point detection, Commun. Stat. Part B Simul. Comput., 32, 319, 10.1081/SAC-120017494 · DOI · Paper page
- Chen, N., Yang, Z., Chen, Y., and Polunchenko, A. (2017, January 1–4). Online anomalous vehicle detection at the edge using multidimensional SSA. Proceedings of the 2017 IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2017, Atlanta, GA, USA. · doi:10.1109/infcomw.2017.8116487
- Aggarwal, C.C. (2017). Outlier Analysis, Springer. · DOI · Paper page
- Li, 2021, Clustering-based anomaly detection in multivariate time series data, Appl. Soft Comput., 100, 106919, 10.1016/j.asoc.2020.106919 · DOI
- Nie, F., Li, Z., Wang, R., and Li, X. (2022). An Effective and Efficient Algorithm for K-means Clustering with New Formulation. IEEE Trans. Knowl. Data Eng. · DOI
- Singh, R.V., and Bhatia, M.P. (2011, January 3–5). Data clustering with modified K-means algorithm. Proceedings of the International Conference on Recent Trends in Information Technology, ICRTIT 2011, Chennai, India. · DOI
- Budiarto, E.H., Erna Permanasari, A., and Fauziati, S. (2019, January 30–31). Unsupervised anomaly detection using K-Means, local outlier factor and one class SVM. Proceedings of the 2019 5th International Conference on Science and Technology, ICST 2019, Yogyakarta, Indonesia. · DOI
- Hussain, 2019, A k-means based co-clustering (kCC) algorithm for sparse, high dimensional data, Expert Syst. Appl., 118, 20, 10.1016/j.eswa.2018.09.006 · DOI
- Chang, C.P., Hsu, W.C., and Liao, I.E. (2019, January 19–21). Anomaly detection for industrial control systems using k-means and convolutional Autoencoder. Proceedings of the 2019 27th International Conference on Software, Telecommunications and Computer Networks, SoftCOM 2019, Split, Croatia. · DOI · Paper page
- Kiss, I., Genge, B., Haller, P., and Sebestyen, G. (2014, January 4–6). Data clustering-based anomaly detection in industrial control systems. Proceedings of the 2014 IEEE 10th International Conference on Intelligent Computer Communication and Processing, ICCP 2014, Cluj-Napoca, Cluj, Romania. · DOI
- Guo, S., Lin, W., Zhao, K., and Su, Y. (2021, January 18–20). Comparison of Clustering-based Network Traffic Anomaly Detection Methods. Proceedings of the IMCEC 2021—IEEE 4th Advanced Information Management, Communicates, Electronic and Automation Control Conference, Hongqing, China. · DOI
- Ester, M., Kriegel, H.P., Sander, J., and Xu, X. (1996, January 2–4). A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise. Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, KDD’96, Portland, OR, USA.
- Liu, 2021, A Real-Time Detection Method for Abnormal Data of Internet of Things Sensors Based on Mobile Edge Computing, Math. Probl. Eng., 2021, 6655346
- Garg, 2020, A multi-stage anomaly detection scheme for augmenting the security in IoT-enabled applications, Future Gener. Comput. Syst., 104, 105, 10.1016/j.future.2019.09.038 · DOI · Paper page
- Audibert, 2022, Do deep neural networks contribute to multivariate time series anomaly detection?, Pattern Recognit., 132, 108945, 10.1016/j.patcog.2022.108945 · DOI
- Li, 2003, Improving one-class SVM for anomaly detection, Int. Conf. Mach. Learn. Cybern., 5, 3077
- Barbariol, T., Chiara, F.D., Marcato, D., and Susto, G.A. (2022). Control Charts and Machine Learning for Anomaly Detection in Manufacturing, Springer.
- Liu, 2012, Isolation-Based Anomaly Detection, ACM Trans. Knowl. Discov. Data, 6, 39, 10.1145/2133360.2133363 · DOI · Paper page
- Li, N., Liu, X., Liu, Z., Mao, L., Zhao, L., and Wang, X. (2021, January 14–17). Anomaly Detection in Power Grid IoT System based on Isolated Forest. Proceedings of the ACM International Conference Proceeding Series, Melbourne, VIC, Australia. · doi:10.1145/3498851.3498922
- Liu, 2021, Sensors Anomaly Detection of Industrial Internet of Things Based on Isolated Forest Algorithm and Data Compression, Sci. Program., 2021, 6699313
- Priyanto, C.Y., and Purnomo, H.D. (2021, January 23–25). Combination of Isolation Forest and LSTM Autoencoder for Anomaly Detectio. Proceedings of the 2021 2nd International Conference on Innovative and Creative Information Technology, ICITech 2021, Salatiga, Indonesia. · DOI
- Hariri, 2021, Extended Isolation Forest, IEEE Trans. Knowl. Data Eng., 33, 1479, 10.1109/TKDE.2019.2947676 · DOI · Paper page
- Cho, K., Van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., and Bengio, Y. (2014). Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation. arXiv. · DOI · Paper page
- Chung, J., Gulcehre, C., Cho, K., and Bengio, Y. (2014). Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling. arXiv.
- Malhotra, P., Vig, L., Shroff, G., and Agarwal, P. (2015, January 22–24). Long short term memory networks for anomaly detection in time series. Proceedings of the The European Symposium on Artificial Neural Networks, Bruges, Belgium.
- Albawi, S., Mohammed, T.A., and Al-Zawi, S. (2018, January 21–23). Understanding of a convolutional neural network. Proceedings of 2017 International Conference on Engineering and Technology, ICET 2017, Antalya, Turkey. · DOI · Paper page
- Guokun, L., Wei-Cheng, C., Yiming, Y., and Hanxiao, L. (2018, January 8–12). Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks. Proceedings of the Modeling Long-and Short-Term Temporal Patterns with Deep Neural Networks, Ann Arbor, MI, USA.
- Munir, 2019, DeepAnT: A Deep Learning Approach for Unsupervised Anomaly Detection in Time Series, IEEE Access, 7, 1991, 10.1109/ACCESS.2018.2886457 · DOI · Paper page
- He, 2019, Temporal Convolutional Networks for Anomaly Detection in Time Series, J. Physics Conf. Ser., 1213, 042050, 10.1088/1742-6596/1213/4/042050 · DOI · Paper page
- Guo, 2018, Multidimensional Time Series Anomaly Detection: A GRU-Based Gaussian Mixture Variational Autoencoder Approach, Proc. Mach. Learn. Res., 95, 97
- Sakurada, M., and Yairi, T. (2014, January 2). Anomaly detection using Autoencoders with nonlinear dimensionality reduction. Proceedings of the ACM International Conference Proceeding Series, Gold Coast, QLD, Australia. · DOI · Paper page
- Hasan, M., Choi, J., Neumann, J., Roy-Chowdhury, A.K., and Davis, L.S. (2016, January 27–30). Learning temporal regularity in video sequences. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA. · DOI · Paper page
- Goodfellow, 2020, Generative adversarial networks, ACM, 63, 139, 10.1145/3422622 · DOI · Paper page
- Bashar, M.A., and Nayak, R. (2020, January 1–4). TAnoGAN: Time Series Anomaly Detection with Generative Adversarial Networks. Proceedings of the 2020 IEEE Symposium Series on Computational Intelligence, SSCI, Canberra, ACT, Australia. · DOI
- Li, 2019, MAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks, Int. Conf. Artif. Neural Netw., 11730, 703
- Kipf, T.N., and Welling, M. (2016). Semi-supervised classification with graph convolutional networks. arXiv.
- Ma, X., Wu, J., Xue, S., Yang, J., Zhou, C., Sheng, Q.Z., Xiong, H., and Akoglu, L. (2021). A Comprehensive Survey on Graph Anomaly Detection with Deep Learning. IEEE Trans. Knowl. Data Eng., 1. · DOI · Paper page
- Zhao, H., Wang, Y., Duan, J., Huang, C., Cao, D., Tong, Y., Xu, B., Bai, J., Tong, J., and Zhang, Q. (2020, January 17–20). Multivariate time-series anomaly detection via graph attention network. Proceedings of the IEEE International Conference on Data Mining, ICDM, Sorrento, Italy. · DOI · Paper page
- Vaswani, A., Brain, G., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., and Polosukhin, I. (2017). Attention is All you Need. arXiv.
- Meng, 2020, Spacecraft Anomaly Detection via Transformer Reconstruction Error, Lect. Notes Electr. Eng., 622, 351, 10.1007/978-981-15-1773-0_28 · DOI
- Tuli, 2022, TranAD: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data, Proc. VLDB Endow., 15, 1201, 10.14778/3514061.3514067 · DOI · Paper page
- Chen, 2022, Learning Graph Structures With Transformer for Multivariate Time-Series Anomaly Detection in IoT, IEEE Internet Things J., 9, 9179, 10.1109/JIOT.2021.3100509 · DOI · Paper page
- Malhotra, P., Ramakrishnan, A., Anand, G., Vig, L., Agarwal, P., and Shroff, G. (2016). LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection. arXiv.
- Park, 2018, A Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-Based Variational Autoencoder, IEEE Robot. Autom. Lett., 3, 1544, 10.1109/LRA.2018.2801475 · DOI · Paper page
- Zhang, 2019, A deep neural network for unsupervised anomaly detection and diagnosis in multivariate time series data, Proc. AAAI Conf. Artif. Intell., 33, 1409
- Shi, X., Chen, Z., Wang, H., Yeung, D.Y., Wong, W.K., and Woo, W.C. (2015). Convolutional LSTM network: A machine learning approach for precipitation nowcasting. arXiv.
- 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.
- Julien, 2020, USAD: UnSupervised Anomaly Detection on Multivariate Time Series, Int. Conf. Knowl. Discov. Data Min., 20, 3395
- Dy, 2018, Deep One-Class Classification, Proceedings of the 35th International Conference on Machine Learning, Volume 80, 4393
- Aditya, P.M., and Nils, O.T. (2016, January 11). SWaT: A water treatment testbed for research and training on ICS security. Proceedings of the 2016 International Workshop on Cyber-Physical Systems for Smart Water Networks (CySWater), Vienna, Austria.
- Wen, T., and Keyes, R. (2019). Time Series Anomaly Detection Using Convolutional Neural Networks and Transfer Learning. arXiv.
- Provotar, O.I., Linder, Y.M., and Veres, M.M. (2019, January 18–20). Unsupervised Anomaly Detection in Time Series Using LSTM-Based Autoencoders. Proceedings of the 2019 IEEE International Conference on Advanced Trends in Information Theory, ATIT, Kyiv, Ukraine. · DOI · Paper page
- Kingma, D.P., and Ba, J.L. (2014). Adam: A Method for Stochastic Optimization. arXiv.
Cited by
Wang, F., Jiang, Y., Zhang, R., Wei, A., Xie, J., & Pang, X. (2025) · Sensors · doi:10.3390/s25010190 · Cited by 119
TE-G-SAGE: Explainable Edge-Aware Graph Neural Networks for Network Intrusion Detection
Luša, R., Pintar, D., & Vranić, M. (2025) · Modelling · doi:10.3390/modelling6040165 · Cited by 15
Detecting cyberattacks based on deep neural network approaches in industrial control systems
Ayas, S., Ayas, M. S., Cavdar, B., & Sahin, A. K. (2025) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2025.104206 · Cited by 4
Metadata
- Authors
- Belay, M. A., Blakseth, S. S., Rasheed, A., & Salvo Rossi, P.
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2023-03-06
- Volume
- 23
- Issue
- 5
- Pages
- 2844
- ISSN
- 1424-8220 (ISSN-L)
- OpenAlex
- W4323314880
- Cited by
- 118
- References
- 16
- FWCI
- 10.37
- Type
- journal article
- Language
- en
- Open access
- gold
- Funders
- Research Council of Norway (318899)
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
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.