近红外光谱结合特征变量筛选方法测定茶汤中的氨基酸含量  

Determination of Amino Acid Content in Tea Infusion using NIR Spectroscopy Combined with Characteristic Variables Selection Methods

作　　者：吴彦红[1] 艾施荣[2] 严霖元[1] 杨红飞[1] 胡琪[1] 

机构地区：[1]江西农业大学工学院,江西南昌330045 [2]江西农业大学软件学院,江西南昌330045

出　　处：《江西农业大学学报》2012年第5期1026-1031,共6页Acta Agriculturae Universitatis Jiangxiensis

基　　金：江西省科技计划项目(20112BBF60019);江西省教育厅科学基金项目(GJJ11081)

摘　　要：采用透射方式获取茶汤的近红外光谱,利用特征变量筛选方法从茶汤的近红外光谱中提取氨基酸光谱信息,建立茶汤中氨基酸含量的快速检测模型。分别利用间隔偏最小二乘法(iPLS)和联合区间偏最小二乘法(siPLS)从茶汤的近红外光谱中提取微弱的氨基酸信息,建立其近红外光谱定量分析模型。结果表明,利用两种方法筛选的特征变量都避开了水的强吸收峰影响,但利用siPLS方法建立的模型性能明显好于iPLS的。最优的siPLS模型对校正集样本的相关系数为0.912,交互验证均方根误差为0.185;用预测集中独立样本检验模型性能,其相关系数为0.887,预测均方根误差为0.202。研究结果可为液体茶饮料中的成分实时快速检测提供参考。The objective of this study was to evaluate the capacity of NIR spectroscopy to rapidly predict the content of amino acid in tea infusion. Transmission mode was used to attain NIR spectroscopy of tea infusion. Interval partial least square （iPLS） and synergy interval partial least square （siPLS） were applied to select the feeble amino acid information from NIR spectroscopy of tea infusion in this study. The optimized characteristic variables were used to develop PLS models. The results show that the selected feature variables based on iPLS and siPLS were not within the range of the strong absorbance for water, but the built model using siPLS had better performance than that of iPLS. The optimal siPLS model was achieved with Rc = 0.912 and RMSECV =0. 185 in the calibration set, Rp =0.912 and RMSEP =0.202 in the prediction set. The attained results can provide a reference for the rapid and realtime determination of the components of in liquid tea drinks.

关 键 词：茶汤 氨基酸 近红外光谱 特征变量筛选 

分 类 号：O657.33[理学—分析化学]