Prediction of Industrial Boiler Energy Efficiency Using Stacking Ensemble Learning

Abstract

Accurate predictions of energy efficiency facilitate the energy-saving renovation of industrial boilers. To improve prediction accuracy, this paper proposes an industrial boiler energy efficiency prediction model based on the stacking ensemble learning method. The base models, including LR, RF, GBDT, XGBoost, and ANN, are established in this study, and then the stacking ensemble learning method is employed to integrate these base models into a strong prediction model. Experimental results indicate that the stacking model outperforms the base models in terms of accuracy and stability.