Accessibility navigation

A data-driven energy performance gap prediction model using machine learning

Yilmaz, D., Tanyer, A. M. and Toker, İ. D. ORCID: (2023) A data-driven energy performance gap prediction model using machine learning. Renewable and Sustainable Energy Reviews, 181. 113318. ISSN 1879-0690

Text - Accepted Version
· Available under License Creative Commons Attribution Non-commercial No Derivatives.
· Please see our End User Agreement before downloading.


It is advisable to refer to the publisher's version if you intend to cite from this work. See Guidance on citing.

To link to this item DOI: 10.1016/j.rser.2023.113318


The energy performance gap is a significant obstacle to the realization of ambitions to mitigate the environmental impact of buildings. Although extensive research has been conducted on the causes, minimization, or the quantifying of the energy performance gap in buildings, comparatively minimal work has been done on raising decision-makers awareness of a potential gap. This paper positions project risks at the core of the gap and proposes an innovative performance gap prediction model focusing on heating and electricity demand in buildings by utilizing the machine learning classification. In this research, the performance gap and project risks of 77 buildings was collected via a web-based survey. The predictive performance of the four machine learning algorithms, namely i) Naive Bayes, ii) k-Nearest Neighbors, iii) Support Vector Machine, and iv) Random Forest, were compared to determine the best model. The results obtained revealed that Naive Bayes was better able to predict the direction of the heating performance gap (72.50%), the negative heating performance gap (71.81%), the positive electricity performance gap (77.08%), and the negative electricity performance gap (83.85%). Furthermore, k-Nearest Neighbors and Support Vector Machine were more accurate to predict the direction of the electricity performance gap (79.00%), and the positive heating performance gap (76.04%).

Item Type:Article
Divisions:Science > School of the Built Environment > Construction Management and Engineering
ID Code:111874


Downloads per month over past year

University Staff: Request a correction | Centaur Editors: Update this record

Page navigation