Geometric Insights for Energy-Efficient Building Design: A Differential Geometry-based Regression Approach

Modeling efficient energy in the design of buildings is an important step for sustainability. Current methodologies are unable to accurately predict the energy requirements. The paper introduces a novel approach “Differential Geometry-Based Regression” (DGR) in order to address this problem. DGR combines both conventional regression analysis and manifold learning techniques in order to establish a system where the correlation among the energy utilization and the building parameters are being explained. The DGR model was trained using the dataset sourced from UCI Machine Learning Repository, the precision and accountability are significantly improved. Thus, it may be used in various applications such as optimization of design towards improved levels of energy efficiency within buildings which shows the way forward for sustainable architecture and construction practices.