摘要
测定了溶解于磷酸缓冲溶液中血糖在8000-5000cm-1波数范围的近红外光谱,用偏最小二乘法建立了测定血糖含量的模型.根据数据特点,提出了按浓度大小选择预测集样本的方法,用来进行最佳模型隐变量个数的选择.结果显示,利用近红外光谱并结合化学计量学方法能够成功地进行血糖的定量分析,方法的测定误差为0.197g/L.
Near infrared spectra (NIR) of glucose in phosphate buffer solution are measured in the range of 8 000 - 5 000 cm-1. Partial least squares method is used to build the regression model of glucose concentration to the NIR spectra. A method to select samples in prediction set is proposed, according to the feature of samples, when cross validation is performed. The results obtained show that NIR spectra coupled with chemometrics can be used in quantitative analysis of glucose. The prediction error is 0.197 g/L.
出处
《山东理工大学学报(自然科学版)》
CAS
2005年第2期95-97,共3页
Journal of Shandong University of Technology:Natural Science Edition