This study investigates the combustion behavior of rice husk using thermogravimetric analysis coupled with decision tree regression. Results indicated that increasing heating rates caused elevated burnout (Tb) and peak temperatures (Tp) while extending the active combustion stage. The optimized decision tree model effectively predicts mass loss, demonstrated by a perfect coefficient of determination (R²) of 1 with a low root mean square error (RMSE) of 0.1993 on the validation set. The model’s robustness suggested its potential for accurate mass loss prediction in rice husk combustion.
Read full abstract