摘要
黄酒中氨基酸态氮和酒精度的常规测定方法步骤繁琐,时效性差。在研究黄酒的近红外光谱和化学计量学的基础上,采用偏最小二乘法建立模型,并以该模型对未知黄酒样品的氨基酸态氮和酒精度含量进行预测,验证模型可靠性。氨基酸态氮建模结果:决定系数为95.19%,均方差为0.041;酒精度建模结果:决定系数为96.82%,均方差为0.26。氨基酸态氮外部检验的预测均方差为0.063;酒精度外部检验的预测均方差为0.20。结果表明该方法应用于黄酒品质的检测,操作简便、快速、准确。
The routine method of determining amino acid nitrogen and alcohol degree in rice wine is complicated and time-consuming. On the basis of near infrared spectroscopy of rice wine and chemical stoichiometry, the model has been established using partial least square. The model was verified by predicting amino acid nitrogen and alcohol degree of the rice wine samples. The determination coefficient （R^2 ） of the amino acid nitrogen model and the alcohol model was 95. 19% and 96. 82, respectively. And the root mean square error of cross validation of each model was 0. 041 and 0. 26%. The predictive root mean square error of external inspection of amino acid nitrogen and alcohol degree was 0. 063 and 0. 20,respectively. The result shows that the near Infrared Spectroscopy technology, which is convenient, rapid and accurate, can be applied in the fast determination of rice wine quality.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2008年第3期361-364,共4页
Computers and Applied Chemistry
基金
国家质量监督检验总局科技项目(2005QK124)
关键词
近红外光谱
黄酒
氨基酸态氮
酒精度
偏最小二乘法
near infrared spectroscopy, rice wine, amino acid nitrogen, alcohol degree, partial least square
