A Comprehensive Approach for Automated Anomaly Detection and Enhancement of EPC Datasets for Decarbonisation

Energy Performance Certificates (EPCs) are pivotal for evaluating building energy efficiency, informing decarbonisation policies, and supporting retrofit. However, errors in EPC datasets compromise their reliability. This study develops an automated anomaly detection framework integrating Machine Learning (ML) and auxiliary datasets to enhance EPC data accuracy at scale. By reviewing validation methods and identifying gaps, this paper proposes a systematic framework for improving data quality assurance, reinforcing EPCs as a key resource for policy and decarbonisation efforts.