Boruta-XGBoost Electricity Theft Detection Based on Features of Electric Energy Parameters

Abstract

Electricity theft detection is critical for the safe and effective development of the electric power system. The existing methods that are used to detect electricity theft rely on historical load data and are considered to have poor timeliness. Their detection results have limited reference to power supply enterprises’ investigation on electricity theft. Therefore, this paper proposes the Boruta-XGBoost power theft detection model based on multiple features of electric energy parameters. The model converts electricity theft detection problem into a multiclass classification problem. First, the features of various electric energy parameters that are collected by the electric energy metering device are used to construct the original dataset. Then, the Boruta algorithm is used to select features and reconstruct the dataset based on the selection results. Finally, the reconstructed dataset is used to train an XGBoost model that can detect the type of electricity theft based on the features of real-time electric energy parameters. In order to verify the effectiveness of the model, a comparative experiment is conducted in this paper. The reconstructed dataset is used to train the following algorithms: LightGBM, CART, SVM, and logistic regression models. The test results show that the Boruta-XGBoost model in this paper has the best effect on electricity theft detection.

Intelligent street lamp control system based on LSTM and Bayesian optimization

“…It has been proved by practice in engineering that it has good fitting ability for various nonlinear models. EXtreme Gradient Boosting (XGBoost) [6] is an integrated learning model based on gradient lifting decision tree. It can generate models with high accuracy by continuously iterating fitting residuals to improve the performance of weak classifiers.…”

Using machine learning ensemble method for detection of energy theft in smart meters

“…The training of the model is conducted in an additive manner [31]. Let $${\hat{y}}_i^{( t )}$$ be the prediction term of the i th instance at the t th iteration.…”

“…The variable T denotes the score of leaf nodes, while the variable w represents the scores assigned to the said leaf nodes. The inclusion of a regularization term provides a valuable advantage in mitigating the issue of overfitting [6,31].…”

“…The optimization objective for the new tree is represented by Equation (19), which is utilized by XGBoost to facilitate the incorporation of custom loss functions. One notable benefit of XGBoost is its incorporation of regularizations into the loss function [6,31]. This inclusion results in the creation of simpler trees and serves as a preventive measure against overfitting.…”

“…XGBoost generates additional trees in order to rectify the remaining errors in the predictions made by the current sequence of trees. The observed outcome is that the model demonstrates a rapid ability to fit into the training dataset, subsequently leading to overfitting [6,[26][27][28][29][30][31]. Achieving a balanced fitting performance is typically accomplished by adjusting the hyperparameters of the machine learning model during both the training and testing phases [6].…”