ANOMALY DETECTION FRAMEWORKS TAXONOMY
DOI:
https://doi.org/10.35363/ViA.sts.2025.142Keywords:
anomaly detection, anomaly detection framework, anomaly detection taxonomyAbstract
In recent years, the increase of digital systems and the consequent need for data volumes have amplified the demand for intelligent systems capable of autonomous monitoring and decision-making. Anomaly detection, the process of identifying patterns in data that deviate significantly from expected behaviour (Chandola, 2009), plays a crucial role in applications such as fault detection, cybersecurity, predictive maintenance, and fraud detection (Nassif A, 2021). However, the increasingly complex nature of modern systems presents challenges in designing anomaly detection frameworks that are scalable and adaptable across various contexts (M. R. Alam, 2019). This research introduces a taxonomy of anomaly detection frameworks, focusing on their capabilities, design dimensions, and levels of abstraction. The proposed taxonomy aims to guide the evaluation of existing frameworks and the design of solutions that are robust, interpretable, and deployable in real-world environments.
References
Chandola, V. B. (2009). Anomaly Detection: A survey. ACM Computing Surveys, 41(3), 1-58. https://doi.org/10.1145/1541880.1541882
M. R. Alam, I. G. (2019). A Framework for Tunable Anomaly Detection. IEEE International Conference on Software Architecture (ICSA), 201-210. https://doi.org/10.1109/ICSA.2019.00029
Nassif A, T. M. (2021). Machine Learning for Anomaly Detection: A Systematic Review. IEEE Access, 9, 78658-78700. https://doi.org/10.1109/ACCESS.2021.3083060
