DATA ANOMALY TAXONOMY AND METHOD SELECTION
Anomaly detection has become a core component of modern data analysis due to the growth of data volumes, the increasing complexity of information systems, and the demand for real-time monitoring. Although anomalies are usually rare, they often correspond to high-impact events such as equipment failures, fraud, cyberattacks, and critical medical conditions. There is no universal notion of an anomaly: deviations can be absolute, context-dependent, or emergent at the level of groups and sequences. Misclassifying the anomaly type leads to inappropriate modeling assumptions, poorly calibrated thresholds, and a trade-off skewed toward false alarms or missed events. This paper reviews three major anomaly types - global (point), contextual, and collective – and relates them to three detection paradigms: statistical tests, density/distance-based methods, and model-based approaches. For each type, we discuss representative algorithms (Z-score, IQR, Mahalanobis distance, kNN/LOF, Isolation Forest, clustering, time-series and sequence models, LSTM/autoencoders, and variational models), together with their data requirements and practical limitations. We provide a method-selection scheme aligned with data modality (tabular data, time series, streaming data), recommend threshold calibration strategies, and outline evaluation protocols for highly imbalanced settings (PR-AUC, MCC, event-based metrics). Correct anomaly typing is a prerequisite for effective monitoring; in applied scenarios, the most robust solutions are typically cascade and ensemble pipelines that combine interpretable baselines with flexible machine learning models.
Kotov D.V. Data Anomaly Taxonomy and Method Selection // Research result. Information technologies. – T.11, №1, 2026. – P. 106-116. DOI: 10.18413/2518-1092-2026-11-1-0-9
While nobody left any comments to this publication.
You can be first.
1. Larin D.O. Information revolutions and their role in the development of mankind / D.O. Larin // Bulletin of Omsk University. – 2025. - Vol. 30, No. 1. – Pp. 37-50. – DOI 10.24147/1812-3996.2025.1.37-50. – EDN GJVYOV.
2. Shkodyrev V.P. Overview of anomaly detection methods in data streams / V.P. Shkodyrev, K.I. Yagafarov, V.A. Bashtovenko, E.E. Ilyina // Proceedings of the Second Conference on Software Engineering and Information Management (SEIM-2017). – St. Petersburg: SPbSUT, 2017. – Vol. 1864. – Pp. 215-225.
3. Shorgin S.Ya. Statistics and clusters in search of anomalous inclusions under big data conditions // Informatics and Its Applications. – 2021. – Vol. 15, No. 4. – Pp. 142-151. – DOI: 10.15393/j12.art.2021.7987.
4. Andrianova E.G., Golovin S.A., Zykov S.V., Les'ko S.A., Chukalina E.R. Review of modern models and methods for analyzing time series dynamics in social, economic, and sociotechnical systems // Russian Technological Journal. – 2020. – Vol. 8, No. 4. – Pp. 7-45. – DOI: 10.32362/2500-316X-2020-8-4-7-45.
5. Vidishcheva E.V., Kopyrin A.S., Vasilenko M.S. Analysis and refinement of anomaly and outlier classification on economic data // Bulletin of the Altai Academy of Economics and Law. – 2019. – No. 6-1. –
Pp. 41-46. – URL: https://vaael.ru/ru/article/view?id=589 (date of access: 22.01.2026).
6. Andrianova E.G., Zykov S.V., et al. Review of modern models and methods for time series analysis // Russian Technological Journal. – 2020. – Vol. 8, No. 4. – Pp. 7-45.
7. Bardasova I.A., Volkova E.A. Anomaly detection in emails using machine learning // Bulletin of Science, No. 5(74), Vol. 4, pp. 1350-1358. 2024. ISSN 2712-8849 // URL: https://www.vesnik-nauki.rf/article/14991 (Accessed: 22.01.2026).
8. Mikhailov A.N. Anomaly detection in network traffic using machine learning methods // Bulletin of Science, No. 12(81), Vol. 3, pp. 1463-1466. 2024. ISSN 2712-8849 // URL: https://www.vesnik-nauki.rf/article/19907 (Accessed: 22.01.2026).
9. Domashkin A.A. Application of two-stage clustering method for anomaly detection: Conference abstract / A.A. Domashkin // International Conference on Computer Systems and Technologies (ICCSS-2024). – Moscow: IPU, 2024. – Pp. 112-120. – URL: https://iccss2024.ipu.ru/proceedings/Домашкин.pdf (accessed: 22.01.2026).
10. Glukhov K.A. Application of two-stage clustering method based on self-organizing Kohonen map for anomaly detection in synthetic datasets / K.A. Glukhov, A.A. Domashkin // Secure and Information Technologies. – 2023. – Vol. 1, No. 1. – Pp. 1-10. – URL: https://info-secur.ru/index.php/ojs/article/view/482 (date of access: 22.01.2026).
11. Kraeva Ya.A. Neural network method for anomaly detection in multidimensional streaming time series // Bulletin of SPbPU. Series: Radio Engineering, Telecommunications and Computer Engineering. – 2024. – No. 2. – Pp. 45-58.
12. Gritsenko A.V. Types of anomalies in video images // Applied Informatics. – 2012. – No. 5. – Pp. 78-92.
13. Litvinovich A.V., Smirnov S.V. Methods for analyzing multidimensional data in anomaly detection tasks // Software Products and Systems. – 2022. – Vol. 135, No. 3. – P. 45-52.
14. Gerasimov M.A., Petrov I.V. Anomaly Detection in Large-Scale Data Using Isolation Forest and an Autoencoder // Bulletin of St. Petersburg State University. Series 15. Computational Mathematics and Informatics. – 2024. – Vol. 20, No. 1. – P. 112–125.
15. Levshun D.A., Popov D.A., Kozlov A.S. Detection and explanation of anomalies in industrial IoT systems based on autoencoders // Software Products and Systems. – 2023. – Vol. 141, No. 4. – P. 123–135.
16. Butusov D.N. Numerical methods for analyzing non-stationary signals in image processing tasks tasks [Doctoral dissertation, Cand. Phys.-Math. Sci., 05.12.04]. SPbGETU "LETİ". – St. Petersburg, 2021. – 150 p.
17. Nosko V.P. Introduction to regression analysis of time series // [study guide]. – Moscow: HSE, 2010. – 120 p.
18. Brykin D.O. Investigation of time series processing algorithms considering non-stationarity [Doctoral dissertation, Cand. Phys.-Math. Sci., 05.13.18]. MIPT. – Moscow, 2023. – 145 p.
19. Lyubushin A.A. Analysis of geophysical and engineering monitoring system data. – 3rd ed. – M.: Nauka, 2024. – 320 p.