Electricity Theft Detection Based on Bagging Heterogeneous Ensemble Learning

Aiming at the deficiency of the single classification method in traditional electricity theft detection, this paper proposes an electricity theft detection method based on Bagging heterogeneous ensemble learning. Considering the performance of different individual learners on the data sets and the diversity between various learners, an electricity theft detection model based on Bagging heterogeneous ensemble learning with combination of various individual learners is developed. The individual learners of the model include the k-nearest neighbors, the error back propagation network, the gradient boosting decision tree and random forest. The outputs of the individual learners are combined by an improved weighted voting strategy. The Irish smart meter data set is used to verify the feasibility of the algorithm. The results show that, compared with the traditional single model, the electricity theft detection based on Bagging ensemble learning has better performance in accuracy, true positive rate and false positive rate. The sensitivity analysis shows the validity of the electricity theft detection method based on Bagging heterogeneous ensemble learning. © 2021 Automation of Electric Power Systems Press.