Assessing the relationship between electricity theft and customer payment habits: A machine-learning approach

IF 3.8 3区经济学 Q3 ENERGY & FUELS

Utilities Policy Pub Date : 2025-04-30 DOI:10.1016/j.jup.2025.101955

Emine Ceren Özay, Erman Çakıt

{"title":"Assessing the relationship between electricity theft and customer payment habits: A machine-learning approach","authors":"Emine Ceren Özay, Erman Çakıt","doi":"10.1016/j.jup.2025.101955","DOIUrl":null,"url":null,"abstract":"<div><div>This study applies machine learning models to historical data to predict the likelihood of debt repayment among customers previously identified as unlawful electricity users. Key predictive factors include past payment behavior, debt levels, demographic characteristics, and regional electricity leakage rates. While existing research on electricity theft predominantly emphasizes detection, repayment prediction remains underexplored. Moreover, although payment behavior analysis is well-established in the financial sector, its application in electricity is limited. This study addresses this gap by demonstrating the effectiveness of Gradient Boosting Machines (GBMs), which outperformed all other models with an accuracy of 90.98 %, an F1 score of 92.20, and an AUC of 0.91. These results highlight GBMs’ capability to manage imbalanced data and capture complex, nonlinear relationships. The findings offer practical value for electricity distribution companies in Türkiye by enabling proactive risk assessment and targeted revenue recovery strategies.</div></div>","PeriodicalId":23554,"journal":{"name":"Utilities Policy","volume":"95 ","pages":"Article 101955"},"PeriodicalIF":3.8000,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Utilities Policy","FirstCategoryId":"96","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0957178725000700","RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENERGY & FUELS","Score":null,"Total":0}

引用次数: 0

Abstract

This study applies machine learning models to historical data to predict the likelihood of debt repayment among customers previously identified as unlawful electricity users. Key predictive factors include past payment behavior, debt levels, demographic characteristics, and regional electricity leakage rates. While existing research on electricity theft predominantly emphasizes detection, repayment prediction remains underexplored. Moreover, although payment behavior analysis is well-established in the financial sector, its application in electricity is limited. This study addresses this gap by demonstrating the effectiveness of Gradient Boosting Machines (GBMs), which outperformed all other models with an accuracy of 90.98 %, an F1 score of 92.20, and an AUC of 0.91. These results highlight GBMs’ capability to manage imbalanced data and capture complex, nonlinear relationships. The findings offer practical value for electricity distribution companies in Türkiye by enabling proactive risk assessment and targeted revenue recovery strategies.

查看原文本刊更多论文

评估电力盗窃与客户支付习惯之间的关系：一种机器学习方法

本研究将机器学习模型应用于历史数据，以预测以前被确定为非法电力用户的客户偿还债务的可能性。关键的预测因素包括过去的支付行为、债务水平、人口特征和地区漏电率。虽然现有的电力盗窃研究主要强调检测，但还款预测仍未得到充分探讨。此外，尽管支付行为分析在金融领域已经很成熟，但在电力领域的应用还很有限。本研究通过证明梯度增强机（GBMs）的有效性来解决这一差距，该模型的准确率为90.98%，F1分数为92.20，AUC为0.91，优于所有其他模型。这些结果突出了GBMs管理不平衡数据和捕获复杂非线性关系的能力。该研究结果通过实现前瞻性风险评估和有针对性的收入恢复策略，为 rkiye配电公司提供了实用价值。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Utilities Policy ENERGY & FUELS-ENVIRONMENTAL SCIENCES

CiteScore

6.80

自引率

10.00%

发文量

审稿时长

66 days

期刊介绍： Utilities Policy is deliberately international, interdisciplinary, and intersectoral. Articles address utility trends and issues in both developed and developing economies. Authors and reviewers come from various disciplines, including economics, political science, sociology, law, finance, accounting, management, and engineering. Areas of focus include the utility and network industries providing essential electricity, natural gas, water and wastewater, solid waste, communications, broadband, postal, and public transportation services. Utilities Policy invites submissions that apply various quantitative and qualitative methods. Contributions are welcome from both established and emerging scholars as well as accomplished practitioners. Interdisciplinary, comparative, and applied works are encouraged. Submissions to the journal should have a clear focus on governance, performance, and/or analysis of public utilities with an aim toward informing the policymaking process and providing recommendations as appropriate. Relevant topics and issues include but are not limited to industry structures and ownership, market design and dynamics, economic development, resource planning, system modeling, accounting and finance, infrastructure investment, supply and demand efficiency, strategic management and productivity, network operations and integration, supply chains, adaptation and flexibility, service-quality standards, benchmarking and metrics, benefit-cost analysis, behavior and incentives, pricing and demand response, economic and environmental regulation, regulatory performance and impact, restructuring and deregulation, and policy institutions.