<![CDATA[Indonesia's heavy reliance on fossil fuels, which account for approximately 80% of its national energy supply, poses a significant obstacle to achieving Net Zero Emissions (NZE) by 2060. Metal-Organic Frameworks (MOF) have emerged as promising innovative materials for sustainable energy applications; however, their limited thermal stability at elevated temperatures remains a major challenge. This study aims to develop a Multilayer Perceptron (MLP -based predictive model for the thermal stability of zinc-based MOF (Zn-MOF) using four structural descriptors nZn, nN, Lig, and Het derived from a dataset of 151 Zn-MOF compounds. Three hidden-layer configurations with 3, 6, and 9 neurons were evaluated using 10-fold cross-validation and three regression metrics: Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and the coefficient of determination (R²). The 9-neuron configuration achieved the highest predictive accuracy, with MAE = 0.0020, RMSE = 0.0022, and R² = 0.9991. SHAP analysis identified nN and Het as the most influential descriptors for thermal stability prediction. These results demonstrate that the MLP architecture effectively captures nonlinear structure–property relationships in Zn-MOFs, offering a computationally efficient tool to accelerate the design of thermally stable materials for sustainable energy applications.]]>
Copyrights © 2026
