Ensemble learning for energy performance prediction of residential buildings
View/ Open
Author
Olu-Ajayi, Razak
Alaka, Hafiz
Sulaimon, Ismail
Grishikashvili, Ketty
Sunmola, Funlade
Oseghale, Raphael
Ajayi, Saheed
Attention
2299/27095
Abstract
In the past decades, the demand for energy in buildings has considerably amplified due to the increase in population and prompt urbanization. The high proportion of energy consumed by buildings engender major environmental problems causing climate change, air pollution and thermal pollution, which is detrimental to human existence. Therefore, the demand for understanding building energy efficiency and how it can be managed effectively is high within academics and society. This elevating concern has increasingly received attention and has been investigated from different perspectives using diverse machine learning techniques such as Support Vector Machine (SVM), Artificial Neural Network(ANN), Decision Tree (DT), among others. There have been applications of ML regression models for the prediction of energy consumption of operational buildings. However, the expedition to develop are liable and accurate model remains elusive. Machine learning classification models can also contributeto bringing more insight to study this issue. In this research, the ensemble learning classification-based method was applied to predict the energy performance of residential buildings. Based on the United Kingdom (UK) Energy Performance Certificate (EPC) standard rating scale, this paper developed and compared six machine learning classification models, namely Support Vector Machine (SVM),Random Forest (RF), Gradient Boosting (GB), Decision Tree (DT), K Nearest Neighbour (KNN) and Extra Trees (ET) for the prediction of building energy performance in terms of performance, feature importance, parameter optimization and computational efficiency. This result shows that ensemble learning produces good results for predicting the energy performance of buildings.