Differentially Private Anomaly Detection for Energy Market Data

• Ana Paixão UNICAMP
• Breno da Silva CPqD
• Filipe H. Cardoso CPqD
• Alexandre Braga CPqD

DOI: https://doi.org/10.5753/sbseg_estendido.2025.11743

Resumo

Differential privacy (DP) is often used to protect consumer privacy in controlled data disclosures. This paper applies DP technology in anomaly detection to safeguard the secrecy of business-sensitive data in Brazilian open energy markets, where full metering datasets are publicly disclosed. Anomaly detection on actual metering data can be found wrong when random noise preserves secrecy, but reduces data utility. This paper contributes to solve this real-world issue by demonstrating DP’s utility for anomaly detection in energy consumption data. We evaluated this approach across various values for the privacy parameter (ϵ), analyzing Precision, Recall, and F1-Score metrics. Our findings facilitate fine-tuning the trade-off between anomaly detection and business secrecy, minimizing the risk of inaccurate insights from noisy data while ensuring robust privacy.

Referências

CCEE (2024). White paper - ccee. [link]. Accessed 2024-11-18.

Dwork, C. (2006). Differential privacy. pages 1–12.

Dwork, C., Kenthapadi, K., McSherry, F., Mironov, I., and Naor, M. (2006a). Our data, ourselves: Privacy via distributed noise generation. pages 486–503.

Dwork, C., McSherry, F., Nissim, K., and Smith, A. (2006b). Calibrating noise to sensitivity in private data analysis. pages 265–284.

ENISA (2023). Cybersecurity and privacy in ai – forecasting demand on electricity grids. Technical report. ENISA.

Fan, L., Xiong, L., and Sunderam, V. (2013). Differentially private anomaly detection with a case study on epidemic outbreak detection. In 2013 IEEE 13th International Conference on Data Mining Workshops (ICDMW), pages 833–840.

Holohan, N., Braghin, S., Aonghusa, P. M., and Levacher, K. (2019). Diffprivlib: the ibm differential privacy library. ArXiv e-prints, 1907.02444 [cs.CR].

Janghyun, K., Barry, H., Tianzhen, H., et al. (2022). A review of preserving privacy in data collected from buildings with differential privacy. Journal of Building Engineering, 56:104724.

Leukam Lako, F., Lajoie-Mazenc, P., and Laurent, M. (2021). Privacy-preserving publication of time-series data in smart grid. Security and Communication Networks, 2021(1):6643566.

Marks, J., Montano, B., Chong, J., Raavi, M., Islam, R., Cerny, T., and Shin, D. (2021). Differential privacy applied to smart meters: A mapping study. pages 761–770. Association for Computing Machinery.

McElroy, T., Roy, A., and Hore, G. (2023). Flip: a utility preserving privacy mechanism for time series. Journal of Machine Learning Research, 24(111):1–29.

Near, J. P., Darais, D., Lefkovitz, N., Howarth, G., et al. (2023). Guidelines for evaluating differential privacy guarantees (nist sp 800-226). Technical report, National Institute of Standards and Technology.

Paixão, A. C. P., Camargo, G. d. F. L., and Braga, A. M. (2025a). Testing open-source libraries for private counts and averages on energy metering time series. In Proceedings of the 2025 20th European Dependable Computing Conference (EDCC), pages 100–104. IEEE.

Paixão, A. C. P., da Silva, B. R., Silva, R. L., Cardoso, F. H., and Braga, A. (2025b). Understanding how to use open-source libraries for differentially private statistics on energy metering time series. In Proceedings of the 10th International Conference on Internet of Things, Big Data and Security (IoTBDSC 2025), pages 289–296. SCITEPRESS - Science and Technology Publications.

Peralta-Peterson, M. and Kotevska, O. (2021). Effectiveness of privacy techniques in smart metering systems. pages 675–678.

Roman, A.-S. (2023). Evaluating the privacy and utility of time-series data perturbation algorithms. Mathematics, 11(5):1260.

Roman, A.-S., Genge, B., Duka, A.-V., and Haller, P. (2021). Privacy-preserving tampering detection in automotive systems. Electronics, 10(24):3161.

Shaham, S., Ghinita, G., Krishnamachari, B., and Shahabi, C. (2024). Differentially private publication of electricity time series data in smart grids. arXiv preprint arXiv:2408.16017.

Zhao, J., Jung, T., Wang, Y., and Li, X. (2014). Achieving differential privacy of data disclosure in the smart grid. pages 504–512.