短波近红外光谱-偏最小二乘法测定白酒中乙醇含量  

Determination of ethanol content in distilled spirits using short-wave near infrared spectroscopy with partial least square regression model

作　　者：王贞佐[1] 逯家辉[1] 郭伟良[1] 陈淑彦[1] 韦韦[1] 滕利荣[1] 

机构地区：[1]吉林大学生命科学学院,吉林长春130012

出　　处：《食品工业科技》2005年第11期163-165,共3页Science and Technology of Food Industry

基　　金：吉林省科技发展基金(批准号:20020503-2)。

摘　　要：应用近红外光谱技术结合偏最小二乘法(NIRS-PLS)建立白酒中乙醇含量定量分析数学模型。所建校正模型相关系数(Corr.Coeff.)达到0.99986,校正集均方根误差(RMSEC)为0.00225,预测均方根误差(RMSEP)为0.00137,模型通过交互验证检验,得出PLS因子数为4时预测残差平方和(PRESS)和交互验证均方根误差(RMSECV)最小。用所建模型测定样品与气相色谱分析结果相对误差不大于0.81%。实验结果表明该方法准确性、稳定性好、精密度高。We have applied near infrared spectroscopy combined with partial least square （NIRS-PLS） to establish a calibration model for determination of ethanol content in Distilled Spirits. In this model, the correlation coefficient （R） in ethanol determination is 0.99986, the root mean square error of calibration （RMSEC） is 0.00225, the root mean square error of prediction （RMSEP） is 0.00137. The model has been verified by crossvalidation. The predicted residual error sum of squares （PRESS） and root mean square error of cross-validation （RMSECV） are the minimum. Compared the results determination by NIRS-PLS with that by gas chromatography （GC） in distilled spirits sample, the relative error is not more than 0.81%. The results of the experiment demonstrate that the accuracy, precision and stability of this method are satisfactory.

关 键 词：近红外 偏最小二乘法 交互验证法 

分 类 号：TS261.7[轻工技术与工程—发酵工程]