包装与食品机械2025，Vol.43Issue(2)：66-74,9.DOI:10.3969/j.issn.1005-1295.2025.02.008
响应面法结合深度神经网络优化刺五加果多糖提取工艺
Optimization of polysaccharide extraction from acanthopanax senticosus fruits using response surface methodology and deep neural network
摘要
Abstract
To enhance the extraction efficiency of polysaccharides from acanthopanax senticosus fruits,this study proposes a microwave-assisted ionic liquid extraction process through a synergistic optimization strategy combining Box-Behnken response surface methodology(RSM)and a deep neural network(DNN).Key parameters(microwave power,ionic liquid concentration,extraction time,and solid-liquid ratio)were screened via RSM experimental design,followed by the establishment of a quadratic regression model.The DNN was subsequently employed to decipher nonlinear interactions among multiple variables and refine process conditions.Results demonstrated that the DNN-optimized conditions(microwave power 350 W,ionic liquid concentration 0.6 mol/L,extraction time 35 min,solid-liquid ratio 1∶24(g/mL)achieved a polysaccharide yield of 16.71%,surpassing the RSM-optimized results.In vitro antioxidant assays revealed IC50 values of 2.36,2.05,and 2.47 mg/mL against hydroxyl radicals,DPPH radicals,and ABTS⁺·radicals,respectively.This work provides a theoretical foundation for developing functional foods and anti-aging nutraceuticals from acanthopanax senticosus fruits.关键词
刺五加果/多糖/工艺优化/响应面法/深度神经网络/抗氧化活性Key words
acanthopanax senticosus fruits/polysaccharides/process optimization/response surface methodology/deep neural network/antioxidant activity分类
轻工业引用本文复制引用
苏适,董立强,黎莉,王双侠,王喜庆,张金凤..响应面法结合深度神经网络优化刺五加果多糖提取工艺[J].包装与食品机械,2025,43(2):66-74,9.基金项目
黑龙江省自然科学基金项目(LH2023H055) （LH2023H055）
黑龙江省本科高校基本科研业务费项目(YWF10236240125) （YWF10236240125）
绥化学院药食同源特色产品开发与应用创新团队项目(240107100305) （240107100305）
