Artificial neural network modeling for impingement drying of <i>Mitragyna speciosa</i> and its antioxidant activity, mitragynine and protein contents

Abstract Mitragyna speciosa (kratom) has gained increasing scientific attention due to its medicinal potential and use in processed herbal products. Drying is essential for extending shelf life and maintaining quality, yet conventional methods often require long durations and high energy, leading to quality degradation. This study aimed to model the drying kinetics of hot air impingement drying (HAID) and microwave-assisted HAID (HAMD) and to evaluate the effects of drying parameters on energy efficiency and product quality. An ANN using a TANSIGMOID transfer function with 10 neurons accurately predicted moisture ratio, drying rate, and moisture content. Higher temperatures shortened drying time and reduced energy use. HAID at 110 °C provided optimal quality, yielding higher chlorophyll retention, greener color, and enhanced phytochemicals and antioxidant activities. HAMD at 110 °C showed the highest energy efficiency, reducing energy consumption by 52.94 % compared to HAID. Protein and mitragynine contents were unaffected, supporting HAMD as a sustainable industrial alternative.