Tian, Zhichao;
Shi, Xing;
Wei, Shen;
(2025)
Exploring the impacts of numerosity reduction on the data-driven building energy analysis for a proposed building.
Journal of Building Performance Simulation
10.1080/19401493.2025.2461256.
(In press).
![[thumbnail of Wei_Exploring the impacts of data reduction on the perfommance with authors info1207.pdf]](https://discovery.ucl.ac.uk/style/images/fileicons/text.png) |
Text
Wei_Exploring the impacts of data reduction on the perfommance with authors info1207.pdf Access restricted to UCL open access staff until 14 February 2026. Download (1MB) |
Abstract
Increasing building performance data promotes data-driven building energy analysis (DDBEA), such as energy prediction and benchmarking. Numerosity reduction is a strategy to filter buildings similar to the target building. However, the impacts of numerosity reduction on DDBEA have not been explored exclusively. This study tends to unveil the effects of numerosity reduction on DDBEA. Three DDBEA scenarios, i.e. energy rating, prediction, and energy-efficient design, are planned to evaluate the impacts of numerosity reduction. The research experiments are tested on three open datasets, which makes it possible to verify and reproduce. Considering key features in energy rating for each building, the results showed that the median of EUI has significant changes ranging from −19.7% to 45.9%. ML models built with numerosity reduction are less stable than traditional models but contain more designable features and are useful for mining causes of the low energy efficiency of the case buildings.
| Type: | Article |
|---|---|
| Title: | Exploring the impacts of numerosity reduction on the data-driven building energy analysis for a proposed building |
| DOI: | 10.1080/19401493.2025.2461256 |
| Publisher version: | https://doi.org/10.1080/19401493.2025.2461256 |
| Language: | English |
| Additional information: | This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. |
| Keywords: | Machine learning, numerosity reduction, building energy prediction, data preprocessing |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of the Built Environment |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10206206 |
Loading...
Loading...Loading...Loading...Archive Staff Only
 |
View Item |
