Artificial intelligence-based modeling and quality evaluation of turmeric under full-spectrum artificial sunlight drying.

Full-spectrum artificial sunlight drying (FSASLD) was investigated as an innovative drying technology for turmeric dehydration, focusing on drying characteristics, quality preservation, and machine learning (ML)-based modeling. Turmeric slices were dried at power levels of 200-400 W using a controlled artificial sunlight drying system. Increasing power level accelerated moisture removal, with effective moisture diffusivity increasing from 2.54 × 10-10 to 5.07 × 10-10 m2 s-1. Color degradation kinetics varied with the evaluated color parameter and drying condition, and both zero-order and first-order models adequately described color changes during drying. Curcumin retention exceeded 85%, indicating effective preservation under FSASLD conditions. Five ML algorithms were employed to predict moisture ratio from drying variables. Among them, the Gaussian process regression model achieved the highest predictive performance with an average test coefficient of determination of 0.9879. These findings highlight the potential of FSASLD as a controllable alternative to traditional drying.