Comparative Evaluation of Regression Models for Building Energy Efficiency Assessment Based on Heating and Cooling Load Requirements

Authors : Sinan Atıcı, Gürkan Tuna

Pages : 283-303

Doi:10.34186/klujes.1804525

View : 107 | Download : 106

Publication Date : 2025-12-31

Article Type : Research Paper

Abstract :The accurate prediction of heating and cooling loads is a critical prerequisite for designing energy-efficient buildings and reducing their environmental footprint. This study presents a comprehensive comparative analysis of multiple regression models for estimating the energy efficiency of residential buildings based on their architectural parameters. Using the Energy Efficiency dataset, we evaluated the performance of seven distinct modelling approaches: Linear Regression, Decision Tree, Random Forest, Support Vector Regression with a Radial Basis Function kernel, K-Nearest Neighbours, Multi-Layer Perceptron, and Deep Neural Networks. Models were rigorously assessed using Root Mean Square Error, Mean Absolute Error, and the coefficient of determination (R²). The results demonstrate that non-linear machine learning methods significantly outperform traditional linear models. Specifically, the Random Forest and Support Vector Regression models achieved superior predictive accuracy, with RMSE values as low as 0.46 for heating load and 1.53 for cooling load, and R² scores exceeding 0.97. Furthermore, feature importance analysis identified Overall Height and Relative Compactness as the most influential parameters for heating and cooling load predictions, respectively, providing actionable insights for architectural design. This research shows that advanced machine learning models, particularly Random Forest and Support Vector Regression, offer a robust and accurate framework for building energy assessment.
Keywords : Enerji verimliliği, Yük tahmini, Makine öğrenimi, Regresyon modelleri, Sürdürülebilir bina tasarımı

ORIGINAL ARTICLE URL