Machine learning based prediction model for building energy ratings in Ireland
Authors
Amin, Basma
Issue Date
2024-05
Degree
MSc in Business Analytics
Publisher
Dublin Business School
Rights holder
Rights
Abstract
Energy can be defined as the driving force behind our modern society. Building energy consumption has emerged as one of the prime contributors to the total energy consumption, due to urbanisation and a colossal increase in the world population. It has been reported that buildings consume 40% of the global energy consumption, and release 38% of Carbon dioxide (CO2). Governments and policymakers are on the lookout for uncovering advanced methods to prepare nations to control climate change and move towards a more sustainable world [1].
Building Energy Ratings have been in the limelight in this regard, as cities have basically turned into blocks of commercial and residential buildings that require energy to fully function. Researchers have proposed the active use of machine learning and technology to support in the agenda. This study has used an official dataset for Irish Building Energy Ratings and emphasises on commissioning the advantages of machine learning to predict energy ratings in Ireland. This study has applied multiclass classification using Logistic Regression, Random Forest Classifier, XGBoost (XGB), Support Vector Classifier and K Nearest Neighbour to train the model to accurately predict building energy ratings. [2]. After the model was applied, it was observed that the Random Forest Classifier and XGBoost were the most efficient models for the purpose of this study. Results from this study can lay a foundation for future studies in the field on building energy ratings in residential and non-residential areas. Energy upgrades have become increasingly common in Ireland and those effected by a better energy certificate have been keen on delving deeper into the possibilities that the Government has offered in this regard.
Incentives and grants are offered to homeowners and landlords for working towards improving the energy ratings of their respective dwellings. The machine learning model implemented in this study can help individuals gauge the energy ratings of their buildings by plugging in the details and features. It would also give individuals a chance to contribute towards sustainability and efficiently utilising scare energy resources of the planet. This would also save time in assessing the Building Energy Ratings (BER) as it is a detailed process and has a number of formalities before a final rating is reached at