Optimization of α-Glucosidase Inhibitors Recovery from Rainbow Trout Hydrolysate Using GA-BP Neural Network

To enhance the value of by-products from rainbow trout processing, the inhibition rate of α-glucosidase (AGIR) is used as indicators. Optimization of rainbow trout hydrolysate extraction by comparing Response Surface Methodology (RSM) and BP Neural Network (BPNN) models. The RSM results: the AGIR is 55.02%. BPNN models predicted the optimal extraction conditions: 51℃ for temperature,1:2.3 for solid-liquid ratio, 4.15 h for time and 0.2334% for enzyme dosage. The reactions obtained under optimized conditions are as follows: the α-glucosidase inhibitory rate increased to 58.14%. This proves that the BPNN model can simultaneously improve the hydrolysis degree and α-glucosidase inhibition rate of the rainbow trout hydrolysate.