摘要
为实现薰衣草精油品种来源快速鉴别的需要,以新疆伊犁地区主要种植的三个薰衣草精油品种为研究对象,基于傅利叶拉曼光谱分析测试技术,先对光谱散射强度数据多元散射校正(MSC)后进行方差分析,找到光谱数据变异最大的三个波长段。然后以二进制方式设定薰衣草精油三个品种变量后,采用正交信号校正(OSC)方法预处理拉曼强度数据,进而构建出不同品种来源的薰衣草精油正交偏最小二乘法(OPLS-DA)识别模型,C-197(2)、法国蓝、H-701三个品种模型的校正均方根残差(RMSEE)分别为0.1096、0.1508、0.1536,交互验证均方根误差(RMSECV)分别为0.1745、0.2394、0.2357,预测均方根残差(RMSEP)分别为0.1965、0.1852、0.1875,未知样本的识别率达到100%,模型效果稳健且准确。通过拉曼光谱结合OPLS-DA分析方法,能够作为薰衣草精油原料质控的一个快速可行方法。
In order to quickly identify the varieties of lavender essential oil,we studied three varieties of lavender essential oil from Xinjiang Yili area.We propose a fast detection method which uses Fourier transformRaman spectra to capture the large amount of spectral data.This approach allowed the discrimination of lavender oil samples from different sources,and to classify them according to Fourier transform-Raman spectroscopy without prior knowledge of this activity.Using MSC pretreatment methods preprocess the Fourier transform-Raman intensity values in original data matrix.Then the variance analysis is carried out to find out the three wavelengths with the largest difference in spectral data.After setting up three binary variables of different variety,orthogonal signal transformation(OSC)is the optimum pretreatment method,getting rid of wavelength points which has nothing to do with determination varieties of lavender essential oil.Finally,the orthogonal partial least square(OPLS-DA)identification model of lavender essential oil has been constructed.In the Identification analysis model of three varieties lavender essential oil,the residual root mean square(RMSEE)of C-197(2),French Blue,H-701 were 0.1096,0.1508,0.1536 respectively.The root mean square error(RMSECV)was 0.1745,0.2394,and 0.2357,respectively.The prediction results show that the root mean square residuals(RMSEP)are 0.1965,0.1852,0.1875,respectively,and the recognition rate of unknown samples reach 100%,which is robust and accurate.The method of Raman spectroscopy combined with OPLSDA analysis can be used as a fast and feasible method for quality control of lavender essential oil.
作者
唐军
韩凯乐
TANG Jun;HAN KaiLe(Physics and Chemistry Detect Center,Xinjiang University,Urumqi 830046,Xinjiang,China;Perilla beauty Biotechnology Co.,Ltd.,Yining 835000,Xinjiang,China)
出处
《计算机与应用化学》
CAS
北大核心
2018年第8期638-644,共7页
Computers and Applied Chemistry
基金
新疆维吾尔自治区重点研发资助项目(2017B03016)