A Hybrid Stacking Ensemble Model with Multidimensional Features for Electricity Theft Detection: Field Validation in West Java
Received: 28 March 2026 | Revised: 27 April 2026 | Accepted: 11 May 2026 | Online: 22 May 2026
Corresponding author: Budi Sudiarto
Abstract
Non-Technical Losses (NTL) in electricity distribution—arising from consumer-level fraud such as meter tampering, illegal connections, and billing manipulation—constitute a persistent financial burden for utilities operating without Advanced Metering Infrastructure (AMI). Existing detection approaches rely predominantly on rule-based threshold systems that fail under high class imbalance, leaving the vast majority of fraudulent accounts undetected in manual-reading networks. This study formulates NTL detection as a binary supervised classification problem at the individual consumer level, using monthly manual-reading data representative of most Indonesian distribution networks. A hybrid stacking ensemble framework integrates 14 features across four domains: consumer behavior, technical infrastructure, socio-economic indicators, and spatial characteristics. LightGBM and CatBoost serve as base learners, and Logistic Regression acts as the meta-learner. Given a severe class imbalance ratio of 22.8:1, Precision-Recall AUC (PR-AUC) is adopted as the primary evaluation metric. The framework was developed and validated on a large-scale dataset from PT PLN (Persero) West Java that included 2.36 million consumers over a 30-month period. Under 5-fold stratified cross-validation, the ensemble achieved a test PR-AUC of 0.7764 (CV: 0.793±0.011), outperforming all single-model baselines. Field validation across 12,407 consumer inspections confirmed a detection rate of 72.0%, compared to 15.3% under the incumbent rule-based system—a 4.7-fold improvement. To the best of our knowledge, this is the largest field-validated machine learning study for electricity theft detection reported for a non-AMI distribution network.
Keywords:
electricity theft detection, machine learning, non-AMI networks, stacking ensemble, imbalanced classification, precision-recall AUC, West Java, non-technical lossesReferences
"World Energy Outlook 2023 – Analysis," International Energy Agency, Oct. 2023. [Online]. Available: https://www.iea.org/reports/world-energy-outlook-2023.
P. Glauner, J. A. Meira, P. Valtchev, R. State, and F. Bettinger, "The Challenge of Non-Technical Loss Detection Using Artificial Intelligence: A Survey," International Journal of Computational Intelligence Systems, vol. 10, no. 1, pp. 760–775, Jan. 2017.
D. Carr and M. Thomson, "Non-Technical Electricity Losses," Energies, vol. 15, no. 6, Mar. 2022.
S. S. S. R. Depuru, L. Wang, and V. Devabhaktuni, "Electricity theft: Overview, issues, prevention and a smart meter based approach to control theft," Energy Policy, vol. 39, no. 2, pp. 1007–1015, Feb. 2011.
R. Czechowski and A. M. Kosek, "The most frequent energy theft techniques and hazards in present power energy consumption," in 2016 Joint Workshop on Cyber- Physical Security and Resilience in Smart Grids (CPSR-SG), Apr. 2016, pp. 1–7.
N. Calamaro, Y. Beck, R. B. Melech, and D. Shmilovitz, "An Energy-Fraud Detection-System Capable of Distinguishing Frauds from Other Energy Flow Anomalies in an Urban Environment," Sustainability, vol. 13, no. 19, Sept. 2021.
M. H. Sidiq, D. Arla, S. H. Fadliqaf, and Q. Haramaini, "Anti Theft Power Smart Metering System," in 2024 International Conference on Technology and Policy in Energy and Electric Power (ICTPEP), Sept. 2024, pp. 287–290.
H. O. Henriques et al., "Development of adapted ammeter for fraud detection in low-voltage installations," Measurement, vol. 56, pp. 1–7, Oct. 2014.
Z. Yan and H. Wen, "Electricity Theft Detection Base on Extreme Gradient Boosting in AMI," IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1–9, 2021.
L. M. R. Raggi, F. C. L. Trindade, V. C. Cunha, and W. Freitas, "Non-Technical Loss Identification by Using Data Analytics and Customer Smart Meters," IEEE Transactions on Power Delivery, vol. 35, no. 6, pp. 2700–2710, Sept. 2020.
L. Wei, A. Sundararajan, A. I. Sarwat, S. Biswas, and E. Ibrahim, "A distributed intelligent framework for electricity theft detection using benford’s law and stackelberg game," in 2017 Resilience Week (RWS), Sept. 2017, pp. 5–11.
Z. Yan and H. Wen, "Performance Analysis of Electricity Theft Detection for the Smart Grid: An Overview," IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1–28, 2022.
J. Nagi, K. S. Yap, S. K. Tiong, S. K. Ahmed, and M. Mohamad, "Nontechnical Loss Detection for Metered Customers in Power Utility Using Support Vector Machines," IEEE Transactions on Power Delivery, vol. 25, no. 2, pp. 1162–1171, Apr. 2010.
A. Jindal, A. Dua, K. Kaur, M. Singh, N. Kumar, and S. Mishra, "Decision Tree and SVM-Based Data Analytics for Theft Detection in Smart Grid," IEEE Transactions on Industrial Informatics, vol. 12, no. 3, pp. 1005–1016, June 2016.
S. M. Saqib et al., "Deep learning-based electricity theft prediction in non-smart grid environments," Heliyon, vol. 10, no. 15, Aug. 2024.
Y. Maraden et al., "Enhancing Electricity Theft Detection through K-Nearest Neighbors and Logistic Regression Algorithms with Synthetic Minority Oversampling Technique: A Case Study on State Electricity Company (PLN) Customer Data," Energies, vol. 16, no. 14, July 2023.
Y. Wang, M. M. Rosli, N. Musa, and F. Li, "Multi-Class Imbalanced Data Classification: A Systematic Mapping Study," Engineering, Technology & Applied Science Research, vol. 14, no. 3, pp. 14183–14190, June 2024.
B. Nurhuda, "Design and development of an electricity theft detection system using XGBoost," B.S. Thesis, Universitas Indonesia, 2024.
A. S. Manneken, "Electricity theft detection based on machine learning using the logistic regression method," B.S. Thesis, Universitas Indonesia, 2022.
S. M. Saqib et al., "Utilizing machine learning ensembles for effective electricity theft detection," Energy Exploration & Exploitation, vol. 44, no. 1, pp. 526–553, Jan. 2026.
D. Perez, M. Flores, P. Castaneda, J. Santisteban, and A. Onate-Andino, "Detection of Non-Technical Losses in Electrical Metering Systems in Northern Lima Using Predictive Modeling and Business Intelligence," Engineering, Technology & Applied Science Research, vol. 16, no. 1, pp. 31624–31631, Feb. 2026.
M. M. Buzau, J. Tejedor-Aguilera, P. Cruz-Romero, and A. Gómez-Expósito, "Detection of Non-Technical Losses Using Smart Meter Data and Supervised Learning," IEEE Transactions on Smart Grid, vol. 10, no. 3, pp. 2661–2670, Feb. 2019.
Q. Wang, Z. Luo, and P. Li, "Natural Gas Consumption Forecasting Based on Homoheterogeneous Stacking Ensemble Learning," Sustainability, vol. 16, no. 19, Oct. 2024.
H. Iftikhar et al., "Electricity theft detection in smart grid using machine learning," Frontiers in Energy Research, vol. 12, Mar. 2024.
L. Grinsztajn, E. Oyallon, and G. Varoquaux, "Why do tree-based models still outperform deep learning on typical tabular data?," Advances in Neural Information Processing Systems, vol. 35, pp. 507–520, Dec. 2022.
R. Shwartz-Ziv and A. Armon, "Tabular data: Deep learning is not all you need," Information Fusion, vol. 81, pp. 84–90, May 2022.
S. Zian, S. A. Kareem, and K. D. Varathan, "An Empirical Evaluation of Stacked Ensembles With Different Meta-Learners in Imbalanced Classification," IEEE Access, vol. 9, pp. 87434–87452, 2021.
S. Bahrami, E. Yumuk, A. Kerem, B. Topçu, and A. Kaya, "Electricity Theft Detection Using Rule-Based Machine Leaning (rML) Approach," Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji, vol. 12, no. 2, pp. 438–456, June 2024.
Downloads
How to Cite
License
Copyright (c) 2026 Qashtalani Haramaini, Ismi Rosyiana Fitri, Fauzan Hanif Jufri, Iwa Garniwa, Hazlie Mokhlis, Budi Sudiarto
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain the copyright and grant the journal the right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) after its publication in ETASR with an acknowledgement of its initial publication in this journal.
