Open Access Open Access  Restricted Access Subscription Access

Anomaly Detection in Time Series Data: Exploring Algorithms and Methods

Akash Tiwari, Rajan Kumar Maurya

Abstract


In today's data-driven world, the ability to detect anomalies in time series data is paramount for a wide range of applications, from finance to cybersecurity and beyond. This research paper embarks on a journey to explore the realm of anomaly detection in time series data, with a focus on the algorithms and methods employed to uncover hidden irregularities. The literature review sets the stage by examining the existing landscape of time series anomaly detection, offering a comprehensive overview of the methods and challenges faced in this domain. It becomes evident that time series data possess unique characteristics, including seasonality, trends, and cyclic patterns, necessitating specialized techniques for anomaly detection. Our paper delves deep into these specialized techniques, dissecting the inner workings of popular algorithms. Statistical methods like the z-score and modified z-score are examined alongside machine learning techniques such as Isolation Forest and One-Class SVM. We also explore deep learning approaches, particularly those based on Long Short-Term Memory (LSTM) networks. The aim is to provide readers with a solid understanding of each algorithm's principles and suitability for different scenarios. To evaluate the efficacy of these algorithms, we introduce a suite of evaluation metrics, shedding light on their performance. Beyond the algorithms, we address the challenges and considerations involved in time series anomaly detection, from handling imbalanced data to selecting appropriate features. Case studies and practical applications illuminate the paper, demonstrating how these methods can be applied in real-world contexts. From financial markets to cybersecurity incidents, the impact of time series anomaly detection is far-reaching. Looking ahead, we explore emerging trends and future directions in this field, offering insights into where research may lead. In conclusion, this paper serves as a comprehensive guide to understanding and implementing anomaly detection in time series data, empowering researchers and practitioners to uncover hidden insights in their temporal datasets.


Full Text:

PDF

References


Yang, Q., Fu, X., Mahanti, A., & Tang, J. L. (2022). Anomaly Detection in Recommendation Systems: A Survey. ACM Transactions on Intelligent Systems and Technology (TIST), 13(4), 1-36.

Chandola, V., Banerjee, A., & Kumar, V. (2009). Anomaly Detection in Time Series Data: A Survey. IEEE Transactions on Knowledge and Data Engineering, 21(12), 1575-1587.

Zhou, G., Mahanti, A., & Tang, J. L. (2020). Deep Anomaly Detection for Recommender Systems. In Proceedings of the 2020 ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 3781-3789).

Ravì, D., & Tantipathpong, C. (2020). Deep Learning for Anomaly Detection: A Survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 10(4), e1361.

Feng, L., Keogh, E., Hu, W., Xing, Y., & Zhu, Q. (2015). A Deep Neural Network Approach for Anomaly Detection in Time Series. In Proceedings of the 2015 SIAM International Conference on Data Mining (pp. 4-12).

Feng, L., Keogh, E., & Zhu, Q. (2017). A Deep Learning Framework for Anomaly Detection in Time Series Data. In Proceedings of the 2017 IEEE International Conference on Data Mining (pp. 933-942).

Mahanti, A., Zhou, G., & Tang, J. L. (2020). Deep Anomaly Detection Using Generative Adversarial Networks. ACM Transactions on Intelligent Systems and Technology (TIST), 11(3), 1-28.

Mahanti, A., Zhou, G., & Tang, J. L. (2019). Anomaly Detection Using Conditional Variational Autoencoders. arXiv preprint arXiv:1902.03802.

Mahanti, A., Zhou, G., & Tang, J. L. (2018). Deep Anomaly Detection Using Stacked Variational Autoencoders. In Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing (pp. 1095-1106).

Aggarwal, C. C., & Aggarwal, C. C. (2021). A Survey on Anomaly Detection in Time Series Data. IEEE Transactions on Knowledge and Data Engineering, 33(1), 1-34.


Refbacks

  • There are currently no refbacks.