Detection of abnormalities and electricity theft using genetic Support Vector Machines
Detection of abnormalities and electricity theft using genetic Support Vector Machines
复制标题
DOI：
10.1109/tencon.2008.4766403
复制
发表时间：
2008-11
期刊：
影响因子：
--
通讯作者：
J. Nagi;K. S. Yap;S. Tiong;Syed Khaleel Ahmed;A. Mohammad
中科院分区：
文献类型：
--
作者：
J. Nagi;K. S. Yap;S. Tiong;Syed Khaleel Ahmed;A. Mohammad
近年来，检测电力欺诈的有效方法一直是一个活跃的研究领域。本文提出了一种使用遗传算法 (GA) 和支持向量机 (SVM) 对电力公司进行非技术损失 (NTL) 分析的混合方法。这项研究的主要动机是协助马来西亚国家能源有限公司 (TNB) 减少其分销领域的 NTL。这种混合GA-SVM模型根据异常消费行为预先选择可疑客户进行现场检查是否存在欺诈行为。所提出的方法使用客户负载配置文件信息来暴露已知与 NTL 活动高度相关的异常行为。 GA 使用随机和预填充基因组的组合提供增强的收敛性和全局优化的 SVM 超参数。欺诈检测模型的结果产生分类类别，用于筛选潜在的欺诈嫌疑人以进行现场检查。仿真结果证明，与TNB当前为减少NTL活动而采取的行动相比，所提出的方法更加有效。
Efficient methods for detecting electricity fraud has been an active research area in recent years. This paper presents a hybrid approach towards non-technical loss (NTL) analysis for electric utilities using genetic algorithm (GA) and support vector machine (SVM). The main motivation of this study is to assist Tenaga Nasional Berhad (TNB) in Malaysia to reduce its NTLs in the distribution sector. This hybrid GA-SVM model preselects suspected customers to be inspected onsite for fraud based on abnormal consumption behavior. The proposed approach uses customer load profile information to expose abnormal behavior that is known to be highly correlated with NTL activities. GA provides an increased convergence and globally optimized SVM hyper-parameters using a combination of random and prepopulated genomes. The result of the fraud detection model yields classified classes that are used to shortlist potential fraud suspects for onsite inspection. Simulation results prove the proposed method is more effective compared to the current actions taken by TNB in order to reduce NTL activities.