An artificial neural network model for predicting building heating and cooling loads

Abstract

In this work, an artificial neural network is designed for predicting the heating and cooling loads for buildings. The paper develops two neural models that make use of a dataset with 8 input attributes with the output as a numeric value of the heating and cooling loads of the buildings. The predictive abilities of the neural nets are compared with linear regression under conventional validation, 5-fold validation and 10-fold validation. Obtained results show that both neural models obtain encouraging results and can be dependably deployed in building loads determination.