Comparison of Support Vector Machine, Gaussian Process Regression and Decision Tree Models for Energy Consumption Prediction of Campus Buildings

Accurate prediction of campus building energy consumption is an important part of building energy consumption management, which is of great significance for reducing campus energy consumption. This paper compares the prediction accuracy of different machine learning algorithms for campus building energy consumption. Firstly, the meteorological and energy consumption data of a campus building in Beijing were collected. And the data were preprocessed and normalized by interpolation methods. Secondly, based on the collected data, three kinds of energy consumption prediction models of campus buildings, namely support vector machine (SVM), Gaussian process regression (GPR) and decision tree (DT), were trained respectively. The results show that the prediction accuracy of SVM model is 5.79% higher than that of GPR model and 13.93% higher than that of DT model. It is proved that SVM has the highest prediction accuracy and good practical value for energy consumption prediction of campus buildings.