摘要
采用AntarisⅡ傅立叶变换红外光谱分析仪器(Thermo Nicolet)测定了9种杜仲的光谱数据,运用偏最小二乘法(PLS)和主成分分析回归(PCR)分别建立了杜仲中松脂醇二葡萄糖苷(PDG)含量与吸光度变量的近红外光谱定标模型,并对所建模型进行验证。结果表明,2种方法建立的模型精度都较高,其中模型的预测能力强弱指标SSE=2.8333,PRESS=7.2392,可见在近红外光谱下,PCR和PDG都适合对杜仲进行检测研究,这是本文的一个重要结果,为以后杜仲的指标检验研究提供了理论和实验依据;但是,PCR建立的(PDG)定标模型预测精度稍高于偏最小二乘法(PLS)建立的模型,所以文中只对PCR进行模型的建立,PLS方法对杜仲检测分析这个问题有待于以后的进一步讨论和研究。
The grades of Eucommia were very complicated which were deteced at home and abroad. High Performance Liquid Chromatography was used in detecing Eucommia by professor in the past,but these methods were complicated and expensive.At the moment, the spectroscopy was used in detecting Eucommia, and the effect was significant.This article focuses on near infrared spectroscopy detection mechanism and results.Nine kinds of Eucommia were detected using Antafignear-infrared spectroscopy, the model between Eucommia pinoresinol diglucoside contents and absorbance variables was established using partial least squares(PLS) and principal component analysis regression (PCR), and the model was detected. In result,the predicting precision of model set up by PCR and PLS is more well, which model is better strength index:SSE=2.8333 PRESS=7.2392. Therefore the testing of PCR is more suitable for reach on Eucommia using Near-infrared spectroscopy, this is an important advance, it provide the theoretical and experimental basis for future research in Eucommia. But from this paper we can know:the predicting precision of model set up by PCR is higher than that by PLS, therefore, this paper only using the PCR model, this question about the PLS for model remains to be further discussion and study in the future.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2011年第3期288-290,共3页
Computers and Applied Chemistry
基金
国家自然科学基金项目(10664001)
贵州省高层次人才特助项目(TZJF200609)
贵州省优秀青年科技人才项目