A trustworthy by design classification model for building energy retrofit decision support

Improving energy efficiency in residential buildings is critical to combating climate change and reducing greenhouse gas emissions. Retrofitting existing buildings –that are major contributors to energy use– is therefore a key priority, particularly in regions with outdated building stock. Artificial Intelligence (AI) and Machine Learning (ML) can automate retrofit decision-making and find retrofit strategies. However, their implementation faces challenges of data availability, trust and alignment with trustworthiness guidelines, as well as compliance to AI regulations. This paper presents a trustworthy-by-design ML-based decision support framework that recommends energy efficiency strategies for residential buildings using minimal user-accessible inputs. The framework employs Conditional Tabular Generative Adversarial Networks (CTGAN) to augment limited and imbalanced data, while neural network-based multi-label classifier identifies potential combinations of retrofit measures. An Explainable AI (XAI) layer using SHAP is also incorporated to clarify the rationale behind recommendations, validate the model, and guide feature engineering. Two case studies on distinct datasets validate performance and replicability: i) a well-established, large Energy Performance Certificate (EPC) dataset for England and Wales; ii) an imbalanced post-retrofit dataset from Latvia (RETROFIT-LAT). Results demonstrate that the framework can handle diverse data conditions and improve performance up to 53% compared to the baseline model without XAI and synthetic data generation. Overall, the proposed framework provides a novel, user-friendly classification-based solution for building retrofit decision support that incorporates the trustworthiness aspects of transparency, human oversight, data governance, and fairness and aids stakeholders in achieving effective energy efficiency investments while aligning with AI regulation and ethical standards.