Machine learning framework for evaluating energy performance certificate (EPC) effectiveness in real estate: A case study of Turin’s private residential market

The Energy Performance Certificate (EPC) is a key tool for advancing building energy efficiency across Europe. By offering standardized information on a property’s energy use, it shapes buyer and tenant preferences, influencing property values. This data-driven policy analysis assesses the EPC's effectiveness. To assess the impact of EPC on property prices in Turin, Italy, a comprehensive machine learning (ML) framework is employed. This framework includes unsupervised hierarchical clustering and supervised algorithms including Artificial Neural Networks (ANN), k-Nearest Neighbors (k-NN), Support Vector Regression (SVR), Random Forest (RF), and Gradient Boosting Machine (GBM). These techniques facilitate an in-depth analysis of the complex relationships between EPC ratings and property prices. Furthermore, the integration of eXplainable Artificial Intelligence (XAI) enhances the transparency of these models, providing clear insights into how EPC ratings affect prices across different property sub-markets. By demystifying the decision-making processes of complex algorithms, this approach makes the findings more accessible to stakeholders. The flexibility of this framework suggests that it can be applied to other European contexts, offering a valuable tool for policymakers aiming to craft more effective energy efficiency strategies. • EPC evaluated in real estate using hierarchical clustering and ML. • GBM with SHAP and PDP offers transparency on EPC impacts, supporting local policies. • XAI provides insights into ML models, enhancing interpretability for stakeholders. • Flexible framework applicable across Europe for energy efficiency strategies. • Real estate market segmented via clustering for more effective EPC analysis.