DETECTION OF ELECTRICITY METER BYPASS IN NIGERIA USING MACHINE LEARNING

Subject Area: Artificial Intelligence

This study focused on the detection of electricity meter bypass in Nigeria using machine learning algorithms. The aim was to develop an intelligent model with the capacity to detect potential customers involved in energy theft through meter bypass. To achieve this, data as collected from the Enugu Electricity Distribution Company (EEDC), considering customer meter recharge information and then transformed to train Support Vector Machine (SVM) and Artificial Neural Network (ANN) algorithms and generate models for the classification of energy theft. The evaluation of the models considered metrics such as accuracy, True Positive Rate (TPR), and False Negative Rate (FNR). The results demonstrated that both SVM and ANN models achieved high accuracy and TPR, with the SVM model having an accuracy of 93.9% and TPR of 100%, and the ANN model achieving 100% accuracy and TPR of 100%. Comparative analyses with existing models showed that the proposed SVM and ANN models outperformed previous methods in terms of accuracy and TPR. The limitation of the study is that it still requires other investigation to prove that a customer involved in energy theft, while the new system facilitated the process.