Title:
Electrical Network Fraud Detection Using a CNN+LSTM Model

Authors:
Rafael A. Gomez and Felipe A. Llaugel, Dominican Republic

Abstract:

This work focuses on the persistent problem of energy loss in the Dominican Republic and in Latin America, with an emphasis on non-technical losses caused by electricity fraud. Given the seriousness of the problem, this paper proposes the development and implementation of a fraud classification model in electricity distribution networks using Deep Learning, specifically a combined model of a Convolutional Neural Network Distributed in Time (CNN) and Long Short Term Memory (LSTM). The goal is to understand and evaluate current fraud detection techniques, investigate the applicability and efficiency of CNN + LSTM models in fraud detection, and address potential challenges in implementing this model in Latin America. The justification lies in the considerable financial losses generated by electricity fraud.

DOI: http://dx.doi.org/10.51505/ijaemr.2023.8505