Elman网络近红外光谱技术同时测定鲜乳中三种主成分含量  

Determination of three components in raw milk using near infrared spectroscopy by elman model

作　　者：荣菡[1] 刘波平[2] 邓泽元[1] 罗香[2] 

机构地区：[1]南昌大学食品科学与技术国家重点实验室,南昌330047 [2]江西省分析测试中心,南昌330029

出　　处：《食品科技》2008年第11期258-261,共4页Food Science and Technology

基　　金：长江学者和创新团队发展计划(No.IRT0540)。

摘　　要：采用Elman神经网络(反馈神经网络,Recurrent Network)结合近红外光谱技术建立鲜乳中的脂肪、蛋白质、乳糖定量分析模型。用偏最小二乘法(PartialLeast Squares,PLS)将原始数据压缩主成分,取前3个主成分的14个吸收峰值输入Elman网络,网络中间层神经元个数为53。Elman网络模型对样品中3个组分含量的预测决定系数(R2)分别为:0.985、0.951、0.967,表明所建Elman网络预测模型能够较准确预测鲜乳中脂肪、蛋白质和乳糖的含量,从而为近红外光谱的多组分定量分析提供了新思路。The paper introduces an application for NIRS multi-component quantitative analysis by setting up a kind of Recurrent Network（Elman） model. Elman prediction model for fat, protein, lactose in raw milk samples had been established with good veracity. 14 peak value data from 3 principal components straight ahead compressed from original data by PLS were taken as inputs of Elman while 3 predictive targets as outputs. 53 nerve cells were taken as hidden nodes with the lowest error. Its training iteration times supposed as 1000. Predictive correlation coefficient by the model are 0.985, 0.951, 0.967. The results show that Elman using in NIRS is a rapid, effective way of measuring fat, protein, lactose in raw milk, and it also can be used as a new idea in the multi-component quantitative analysis of other samples.

关 键 词：近红外光谱技术 ELMAN网络 偏最小二乘法 多组分 鲜乳 

分 类 号：TS201.2[轻工技术与工程—食品科学]