Science Progress, Год журнала: 2025, Номер 108(2)

Опубликована: Апрель 1, 2025

The shaping and drying of ceramics are a critical yet complex process that directly influences ceramic materials’ final properties performance. Predicting key parameters such as the coefficient plasticity, mass loss during in air at point, moisture is essential for optimizing these processes. This study analyzes dataset clays various chemical compositions to predict reveal most important on producing tiles. data then employed develop compare four advanced machine learning models. models were evaluated using performance metrics determination ( R ²), mean absolute percentage error (MAPE), (MAE), root squared (RMSE). Extreme Gradient Boosting (Gradient Boosting) emerged reliable model, with 0.9871 R², 0.2672 RMSE, 0.2086 MAE, 1.61% MAPE. Support vector regression artificial neural networks also delivered strong performances, while random forest, though competitive, was slightly less accurate. Furthermore, model interpretation methods analysis provided valuable validation predictive capabilities influence input features. techniques processes offer robust toolkit enhancing efficiency, reliability, sustainability materials engineering. It seen Al 2 O 3 levels up 23% had little effect plasticity susceptibility, significant changes occurring above 28%. Fe content found between 1.5% 1.7%, SiO about 62%. findings this decision-support tools manufacturers, raw material suppliers, engineers, enabling more informed selection, reduced waste, improved product consistency across industry.

Язык: Английский