IoT-Based Real-Time Monitoring and Detection of Electricity Theft Using CNN-LSTM and Isolation Forest
Journal of Carbon Sequestration and Climate Engineering · 30.06.2026 · 1 (2) · 1-9
DOI: 10.13074/jcsce.2026.06.262006 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Carbon Sequestration and Climate Engineering, online erschienen am 30.06.2026. Verfasst von Udhayakumar M und 4 weiteren Autor:innen.
Abstract (Originalsprache)
The objective of the project is to provide an IoT-based smart electricity theft detection system using Hybrid CNN-LSTM and Isolation Forest. Existing methods work on high accuracy using offline data and supervised models, but lack in real time IOT integration and the ability to detect unseen theft patterns, which our proposed system addresses. Group 1 addresses electricity theft in traditional meters using a Random Forest and KNN hybrid model that achieves 89% accuracy and 98% recall, reducing reliance on manual inspection. Group 2 proposes a smart detection system combining CNN-LSTM and Isolation Forest with real-time IOT integration. Result: The performance of the proposed system [CNN-LSTM, Isolation Forest] model was compared to the existing [RF - KNN] method. The proposed method [CNN-LSTM, Isolation Forest] achieved an Accuracy of 93%, Precision of 84%, Recall of 98.9%, F1-Score of 91%, and AUC of 99% with significance of
Zitieren
APA
M, U., N, A., G, P., S, J., & S, V. V. (2026). IoT-Based Real-Time Monitoring and Detection of Electricity Theft Using CNN-LSTM and Isolation Forest. Journal of Carbon Sequestration and Climate Engineering, 1(2), 1-9. https://doi.org/10.13074/jcsce.2026.06.262006
Vancouver
M U, N A, G P, S J, S VV. IoT-Based Real-Time Monitoring and Detection of Electricity Theft Using CNN-LSTM and Isolation Forest. Journal of Carbon Sequestration and Climate Engineering. 2026;1(2):1-9. doi:10.13074/jcsce.2026.06.262006
BibTeX
@article{M2026,
author = {M, Udhayakumar and N, Arun and G, Perumal and S, Jeyabharathi and S, Vikram V},
title = {IoT-Based Real-Time Monitoring and Detection of Electricity Theft Using CNN-LSTM and Isolation Forest},
journal = {Journal of Carbon Sequestration and Climate Engineering},
year = {2026},
volume = {1},
number = {2},
pages = {1--9},
doi = {10.13074/jcsce.2026.06.262006}
}
EUVIMED