Machine Learning-Based Prediction of Thermal Conductivity in Bio-Composite Materials

Bio-composite materials are increasingly used in the construction sector due to their low thermal conductivity and potential to improve building energy efficiency. Traditionally, thermal conductivity is evaluated through experimental testing or theoretical modeling, which can be time-consuming and costly. This study investigates the use of machine learning techniques to predict the thermal conductivity of bio-composite materials using 93 experimental samples extracted from the literature. Four regression models—Linear Regression, Ridge Regression, Support Vector Regression, and Random Forest—were trained using the same dataset and evaluated through an independent train–test validation strategy. Among the tested models, Random Forest achieved the best predictive performance, with a coefficient of determination R2 = 0.885 and a root mean square error RMSE = 0.042, demonstrating its ability to capture nonlinear relationships in heterogeneous composite materials. These results highlight the potential of machine learning as a complementary tool for rapid estimation of thermal conductivity in sustainable building materials.