摘要
通过均匀设计实验和二次多项式逐步回归分析得出从芥菜中提取多糖的优化工艺,即提取时间130 min、酶浓度1%、pH4．0、提取温度60℃。在此基础上运用基于均匀设计的人工神经网络构建了工艺参数与芥菜多糖提取率之间的数学模型。在优化的工艺参数下对多糖的提取率进行预测,二次多项式逐步回归的预测值与实测值的相对误差为23．02%,而人工神经网络的预测值与实测值的相对误差仅为4．37%。结果表明,人工神经网络比二次多项式逐步回归分析的预测结果更准确。
In this paper, the optimal parameters were obtained by the uniform design and regression techniques as follows： extracting time 130 min, enzyme concentration 1%, pH 4.0 and extracting temperature 60℃. Based on uniform design method, a prediction model for the experimental parameters and the extraction yield of polysaccharide from Brassica juncea was established by artificial neural networks. Prediction of the extraction yield of polysaccharide from Brassica juncea using methods of regressive analysis and the artificial neural networks was studied, the relative error by regressive analysis method is 23.02 96 while the relative error by artificial neural networks method is 4. 371%. The results showed that the artificial neural networks methods is more reliable than regressive analysis method in predicting polysaccharide extraction yield.
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2006年第2期53-56,共4页
Food and Fermentation Industries
基金
国务院侨办基金(05Q0016)资助项目
关键词
芥菜
多糖
均匀设计
回归分析
BP神经网络
Brassica juncea, polysaccharide, uniform design, regressive analysis, BP neural networks
