Electricity Theft Detection Base on Extreme Gradient Boosting in AMI
Zhongzong Yan, He Wen
Abstract
Electricity theft by cyber attacks to smart meters in advanced metering infrastructure causes financial losses to utilities every year. This paper proposes an electricity theft detector using metering data based on Extreme Gradient Boosting (XGBoost). The metering data are preprocessed, including recover missing or erroneous values and normalization. The classification model based on XGBoost are trained using both benign and malicious samples after data preprocessing. Simulations are done by using the Irish Smart Energy Trails with six attack types. Compared with the support vector machine, propagation neural network, and k-nearest neighbors algorithm, the proposed method can detect electricity theft with either higher accuracy or faster speed.