基于近红外光谱对牛奶中 掺杂尿素判别分析
首发时间:2011-12-12
摘要:采集了40个合格的纯牛奶样品,并配置含有尿素质量浓度范围为1g/L-20g/L之间40个牛奶样品,研究了掺杂尿素牛奶的二维相关红外特性,在此基础上选择4200-4800cm-1为建模区间。采用偏最小二乘法建立定性、定量模型,指出通过判别偏最小二乘法(PLS-DA)可以实现纯牛奶及掺杂尿素牛奶的定性鉴别,判别正确率为100%;掺杂牛奶校正集相关系数R为0.999,交叉验证均方差(RMSECV)为0.242,对未知样品集预测相关系数R达到0.999,预测标准偏差(RMSEP)为0.57,这表明所建模型具有较好的预测效果。
关键词: 近红外红外光谱 牛奶 尿素 掺杂 偏最小二乘判别分析法
For information in English, please click here
Discriminant Analysis of Urea in Milk Based on Near Infrared Spectroscopy
Abstract:40 pure milk samples were collected. 40 adulterated milk samples with added different mass concentration of urea (1-20 g/L)were prepared. The two-dimensional correlation spectroscopy was calculated under the perturbation of adulteration concentration. Based on the characteristics of 2D correlation infrared spectroscopy, the spectra in the range of 4200-4800cm-1 was selected to apply. Then all data were analyzed by PLS. Results showed that a 100% recognition ratio of samples was achieved by partial least square-discriminate analysis (PLS-DA). The correlation coefficient(R) of calibration sets is 0.999, while the root mean square errors of cross validation (RMSECV) is 0.242. The R between the predicted values and actual values is also 0.999, while the root mean square errors of prediction (RMSEP) is 0.57, therefore, the model has good prediction ability. It is important to improve quality of dairy products, to protect the benefit of consumers.
Keywords: Near-infrared spectroscopy Milk Urea Adulteration Partial least square-discriminate analysis
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于近红外光谱对牛奶中 掺杂尿素判别分析
评论
全部评论0/1000