Home // IARIA Congress 2024, The 2024 IARIA Annual Congress on Frontiers in Science, Technology, Services, and Applications // View article

Prediction of Residential Building Energy Star Score: A Case Study of New York City

Authors:
Fan Zhang
Baiyun Chen
Fan Wu
Ling Bai

Keywords: Machine learning, Regression, Data analysis, Model evaluation.

Abstract:
Over the past few years, machine learning algorithms have garnered widespread attention in predicting the Energy Star Score of residential buildings. Traditional forecasting models, relying on software and statistical methods, failed to deliver accurate predictions owing to the intricacies of factors, non-linear relationships, and the noise of the data utilized in prediction. In this paper, we propose to use machine learning algorithms to enhance the performance of the Energy Star Score of residential buildings, due to their capability to capture the complex relationships between various kinds of features. We carefully choose the essential features to construct a regression model capable of accurately predicting the score through feature engineering and selection procedures. Additionally, we unveil the significant factors by ranking the importance of various features. Furthermore, we compare the performances of different machine learning algorithms in prediction and identify the optimal model, Gradient Boosting Regressor (GBR), as the best forecaster of Energy Star Scores for residential buildings in New York City. textcolor{black}{GBR outperforms all other methods, exhibiting the lowest Mean Absolute Error (MAE) of 0.89 and Sum of Squared Errors (SSE) of 6199.90, as well as $R^2$ of 0.9967 and adjusted $R^2$ of 0.9966. The variances for all the metrics in the GBR model are also minimized.} Our study results not only enhance the prediction performance of energy scores but also provide valuable insights for decision-makers involved in retrofitting or constructing similar residential buildings with energy-saving considerations.

Pages: 180 to 186

Copyright: Copyright (c) IARIA, 2024

Publication date: June 30, 2024

Published in: conference

ISBN: 978-1-68558-180-0

Location: Porto, Portugal

Dates: from June 30, 2024 to July 4, 2024