Uncovering the financial impact of energy-efficient building characteristics with eXplainable artificial intelligence

Title data

Konhäuser, Koray ; Werner, Tim:
Uncovering the financial impact of energy-efficient building characteristics with eXplainable artificial intelligence.
In: Applied Energy. Vol. 374 (2024) . - 123960.
ISSN 1872-9118
DOI: https://doi.org/10.1016/j.apenergy.2024.123960

Official URL:

Abstract in another language

The urgency to combat climate change through decarbonization efforts is more crucial than ever. The global building sector is one of the primary contributors to carbon emissions, yet the economic implications of energetic building characteristics of residential buildings remain elusive. This study addresses the intersection of building energy performance, market valuation, and carbon emissions reduction by introducing a novel cluster-based eXplainable Artificial Intelligence (XAI) methodology to uncover the financial impact of energetic building features on property valuation. We combine Energy Performance Certificates (EPC) and property transaction data within the UK and apply two sophisticated machine learning models: XGBoost and CatBoost. To this end, we use hierarchical BIRCH clustering to identify subgroups within our comprehensive dataset and leverage SHapley Additive exPlanations (SHAP), Permutation Feature Importance (PFI), and Partial Dependency Plots (PDP) to reveal nuanced insights into the financial contribution of energetic building characteristics to property valuation. This research contributes to the academic discourse by introducing a cluster-based XAI approach for analyzing energy-related financial incentives in the building sector. Our results suggest that energy-efficient building features lead to significant financial benefits outside of London. The cluster-based approach reveals that carbon emissions are predominantly relevant for medium to large buildings outside of London but have a reversed financial effect within the capital. For larger residential buildings, we find a tendency for high running costs for energy (e.g., lighting costs) to be well reflected in transaction prices. The presented findings underscore the potential economic benefits for targeted energy efficiency improvements and illuminate the pathway towards a low-carbon society by addressing inherent uncertainties surrounding the economic viability of energetic investments, thus fostering informed decision-making and sustainable development.

Further data

Item Type:	Article in a journal
Refereed:	Yes
Keywords:	Building energy efficiency; Explainable artificial intelligence; Energy performance certificates; Energy efficient investments; Hierarchical Clustering
Institutions of the University:	Faculties > Faculty of Law, Business and Economics > Department of Business Administration Research Institutions Research Institutions > Affiliated Institutes Research Institutions > Affiliated Institutes > Branch Business and Information Systems Engineering of Fraunhofer FIT Research Institutions > Affiliated Institutes > FIM Research Center for Information Management
Result of work at the UBT:	No
DDC Subjects:	000 Computer Science, information, general works > 004 Computer science 300 Social sciences > 330 Economics
Date Deposited:	03 Sep 2024 05:43
Last Modified:	03 Sep 2024 05:43
URI:	https://eref.uni-bayreuth.de/id/eprint/90328