Identifying Concept Drift with Supervised Algorithms in Smart Grids

Abstract

Early and correct disclosure of anomality is critical for realiable and secure activity of the smart grid. Machine learning based outlier detection schemes could be helpful when there is no concept drift in the smart grid. Concept drift is considered an unforeseeable change in the distribution of the data in machine learning [4]. In this study, an anomaly detection is formulated as a machine learning problem and a robust detection algorithm is proposed. Numerical results demonstrate effectiveness of the proposed solution to detect anomaly and contingency besides integrity attacks in the smart grid.

Authors

Ayşe Sayin, Mostafa Mohammadpourfard, Mehmet Tahir Sandikkaya

Keywords

Machine learning algorithms, Machine learning, Predictive models, Prediction algorithms, Smart grids, Decision trees, Security

---

Publication Type

Journal Article

---

Digital Object Identifier

https://doi.org/10.1109/GTD49768.2023.00076

---

Full Citation

Sayin, A., Mohammadpourfard, M., & Sandikkaya, M. T. (2023, May). Identifying Concept Drift with Supervised Algorithms in Smart Grids. In 2023 IEEE PES GTD International Conference and Exposition (GTD) (pp. 263-267). IEEE.

View Article