Prediction of Heating Load and Cooling Load of Buildings Using Neural Network

Abstract

Prediction of heating load and cooling load is important for the planning and management of energy systems. Recently, numerous studies have focused on load prediction models associated with a wide range of different methods. In addition to regression analysis, a neural network and random forest have been especially studied. Different studies have shown the success and superiority of the neural network over regression analysis for prediction. In this study, the efficiency of heating load and cooling load prediction using the neural network is compared with the results obtained by the logistic regression and random forest. For the demonstration, the dataset describing 12 different forms of buildings are used. The evaluation of the predictions obtained shows that the neural network provides a higher accuracy than the regression analysis and random forest.