Machine learning approach to predict building energy ratings of dwellings in Ireland

Authors

Laxminarayan Bhat, Rahul

Issue Date

2024-05

Degree

MSc in Data Analytics

Publisher

Dublin Business School

Rights

Abstract

The machine learning models have been employed for the BER prediction of a dwelling in Ireland. Various machine learning models have been run on the BER data by researchers. This research focuses on which machine learning model provides an optimal performance. The research also provides an answer to the framed research questions. The prediction of BER using machine learning models will reduce the time required to measure the significant features of the dwelling as the current system requires 211 features to be measured for each dwelling. After running the models, it was found that logic-based techniques perform better on the BER data. For regression analysis, 35 features are highly significant whereas for classification analysis, 42 features are highly significant.