EFI-DSS: A Machine Learning-based Decision Support System for Energy Efficiency Investments in Buildings

The building sector accounts for approximately 40% of global energy consumption and CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions. Moreover, many existing buildings are over 50 years old, resulting in poor energy performance and an urgent need for retrofitting to meet sustainability goals. This paper presents EFI-DSS (Energy Efficiency Investments - Decision Support System), an advanced platform designed to guide users in making cost-effective energy efficiency investments for their buildings requiring minimal information. At its core, EFI-DSS integrates a physics-informed neural network to generate virtual Energy Performance Certificates (EPCs) with minimal input through the user-friendly Graphical User Interface (GUI), even in the absence of detailed energy data. Additionally, the platform allows users to explore a wide range of energy efficiency measures, viewing their costs and thermal capacities while instantly seeing their impact on the building’s EPC rating and total investment cost. An administrative interface enables authorized users to update available measures—such as modifying prices or adding new materials like upgraded doors or windows—ensuring that end-users always have access to the most relevant and up-to-date options. A real-world case study in Riga, Latvia, demonstrates the tool’s effectiveness in supporting data-driven decision-making. EFI-DSS offers residents and building operators clear cost estimates for possible energy performance upgrades, allowing them to make informed investment decisions with confidence and removing financial risk during the first steps of the procedure.