Electrical Load Forecasting Using Hybrid of Extreme Gradient Boosting and Light Gradient Boosting Machine

Abstract

AbstractEnsemble learning methods have been used to improve performance accuracy through bias-variance trade-off techniques. However, there is still room to improve. This paper proposes an ensemble model to forecast the electrical load behavior based on a hybrid of Extreme Gradient Boosting (XGBoost) and Light gradient boosting machine (LGBM). Extreme gradient boosting (XGBoost), a Light gradient boosting machine (LGBM) and a hybrid of XGBoost and LGBM models are trained, evaluated, and compared. The experiments show that the proposed model outperforms other methods by reducing more than 1% in mean absolute percentage error (MAPE), root mean squared percentage error (RMSPE), and mean absolute error (MAE). The dataset from the Pennsylvania-New Jersey-Maryland interconnection power grid was used to validate the evolutionary capability of the proposed method and the finding of optimal accuracy of the model.KeywordsElectrical load forecastingEnsemble learningExtreme gradient boosting machine (XGBoost)Light gradient boosting machine (LGBM)