基于神经网络法制备牡蛎呈味肽工艺优化研究  

Study on the processing optimization of taste peptide enzymatic-production from oyster base on a neural network method

作　　者：侯清娥[1] 秦小明[1] 林华娟[1] 章超桦[1] 刘慧[1] 尤久勇[1] 

机构地区：[1]广东海洋大学食品科技学院,广东湛江524025

出　　处：《食品工业科技》2011年第11期301-304,共4页Science and Technology of Food Industry

基　　金：农业部948项目(2006-G42);现代农业产业技术体系建设专项(贝类;47)

摘　　要：为获得呈味较好且含肽量较高的牡蛎酶解液,选取新鲜牡蛎肉为原料,以酶解液的肽比例、感官分值为指标,采用正交实验,获得酶解时间、温度、加酶量和料水比四因素与肽比例以及感官分值之间的相关性,然后应用神经网络法以此四因素为网络输入,肽比例及感官分值分别为网络输出对酶解工艺分析模拟,并进一步通过遗传算法对酶解工艺进行优化。神经网络系统获得的优化工艺条件:酶解时间5.4h,酶解温度58.6℃,加酶量1.03%,料水比1∶2.8,其呈味肽比例预测值达80.81%,验证实验结果为78.35%;优化工艺条件:酶解时间6.0h,酶解温度53.8℃,加酶量0.95%,料水比1∶2.1,其感官评定预测值为6.67分,验证实验结果为6.39分。以上优化工艺条件,其预测值与验证值的相对误差均在5%以内,表明以神经网络法优化牡蛎酶解工艺条件,可以获得较为理想的目标值。In order to obtain oyster hydrolysis product with good flavor and rich in peptide,enzymatic process optimization was investigated by using flesh oyster as material based on a neutral network method,which built by some parameters obtained from orthogonal experiments（enzymatic time,enzymatic temperature,enzyme concentration,ratio of material to water,peptide concentration and sensory score）.The optimized operation condition obtained by the neutral network method was：enzymatic time 5.4h,enzymatic temperature 58.6℃,enzyme concentration 1.03%,ratio of oyster muscle to water 1∶ 2.8,and its predictive peptide ratio reached to 80.81%,while the result of its confirmatory experiment was 78.35%.The another optimized operation condition was：enzymatic time 6.0h,enzymatic temperature 53.8℃,enzyme concentration 0.95%,ratio of oyster muscle to water 1∶ 2.1,and its predict sensory scores was 6.67,while the result of its confirmatory experiment was 6.39.The relative errors between predict value and confirmatory experiment were all less than 5%,indicated that the neutral network method was an effective tool for optimizing enzymatic process condition for oyster to obtain desired target value.

关 键 词：牡蛎 呈味肽 BP神经网络 遗传算法 

分 类 号：TS254.1[轻工技术与工程—水产品加工及贮藏工程]