Work

Leveraging machine learning for data-driven building energy rate prediction

Year: 2025

Type: article

Abstract: This paper presents a data-driven method for predicting Building Energy Ratings (BER) in urban areas using advanced Machine Learning (ML) algorithms. Focusing on Dublin, it combines geospatial data with building-specific and neighborhood-scale features to classify BER. The study employs Decision Trees, Random Forest (RF), K-Nearest Neighbors, and Support Vector Machines, assessing their performance through metrics such as Receiver Operating Characteristic (ROC), Area Under the Curve (AUC), precision, recall, and overall accuracy. Results identify the RF model as the most effective, achieving AUC values of 0.77–0.83 and excelling in identifying high-energy-consuming buildings. In contrast, the SVM demonstrated limited utility, with AUC values of 0.37–0.49. This research advances Urban Building Energy Modeling (UBEM) by providing a robust tool for policymakers to optimize energy use and improve predictive accuracy. By aligning BER prediction with European energy performance standards (ISO/CEN), it ensures adaptability and relevance across various urban contexts.

Source: Results in Engineering

DOI: 10.1016/j.rineng.2025.104931

Authors: Nasim Eslamirad, Mehdi Golamnia, Payam Sajadi, Francesco Pilla

Institutions: University College Dublin, Tallinn University of Technology

Language: English

Topic: Building Energy and Comfort Optimization

Subfield: Building and Construction

Field: Engineering

Domain: Physical Sciences

SDGs: Sustainable cities and communities, Affordable and clean energy, Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $1,940