Employing Explainable AI to Optimise Domestic Energy for a Greener Society

The pressing global challenge of achieving sustainable energy objectives necessitates innovative approaches to optimise energy consumption in domestic energy supply settings. AI-based expert systems predict energy consumption, aiding energy policy design. However, the inherent complexity and opacity of AI-based systems lead to lack of transparency on prediction mechanisms resulting in resistance in user engagement and decision-making adoption. To bridge this gap, we present here explainable AI-based techniques that enhance AI-based expert systems to provide comprehensible explanation behind the prediction of domestic energy usage and site energy usage intensity (Site EUI) using a dataset of 100 K building energy records collected over seven years in various US states including building characteristics, weather data, and energy consumption. For this purpose, we employ XGBoost for precise Site EUI prediction. Our findings emphasise energy-conscious designs and climate-responsive practices for efficiency. Integrated modelling provides actionable insights for managers and policymakers, driving sustainable energy strategies.