Optimizing the rice wine process of Polygonatum sibiricum Red. using response surface methodology (RSM) and back-propagation (BP) neural network to enhance biological activity and reduce 5-hydroxymethylfurfural (5-HMF)

Traditional processing of Polygonatum sibiricum Red. (PSR), which is rich in polysaccharides, involves prolonged heating at high temperatures and is prone to produce the harmful substance 5-hydroxymethylfurfural (5-HMF). To reduce its generation and improve the taste, the process of wine-processed (PSW) was studied through Response Surface Method and back-propagation neural network model optimized by the sparrow search algorithm as an assessment indicator of the effects of steaming time, drying temperature, drying time and steaming cycles on the contents of polysaccharides, reducing sugars, flavonoids, saponins, 5-HMF and antioxidant activity. The results show that the optimal process is steaming time of 3.6 h, drying temperature of 75 °C, drying time of 11.4 h, and steaming cycles of 8 times, which improved taste and active, and reduced 5-HMF's to 0.18 mg/g. This study provides a reference of PSW processing technology, other heat-processed food and herbs rich in polysaccharides on quality control with safety warning.