Rapid Valuation And Variety Identification In Quality Of Forsythia Suspensa

Near infrared spectroscopy analysis is one of fast developing technology which is nondestructive, fast, efficient, convenient, and has high performance in the detection and analysis of Tradition Chinese Medicine's internal quality. The thesis explores the application of near-infrared spectroscopy in Forsythia suspensa's quality control, and discusses qualitative calibration and quantitative calibration. The work can be summarized as follow:In the aspect of qualitative calibration, Identification method of Forsythia suspense from 5 different habitats by near infrared spectroscopy coupled with pattern recognition based on SIMCA was proposed in this paper. In the spectra region between 8500cm-1 and 4000cm-1, 5 predictive models of Lushi, Luanchuan, Luoning, Xin'an in Henan province, and Anzhe in Shanxi province were built separately by the standard normal variate (SNV) preprocessing method with SIMCA pattern recognition method. optimal principal components were 3, 4, 2, 3, 4. Under the =5% significance level, 1 sample wasn't identified correctly in 10 prediction sample, it prove that method was established can identify exactly Forsythia suspense from different habitats.In the aspect of quantitative calibration, 50 samples were selected , 40 of the samples were selected as calibration set, 10 samples as prediction set. Taking content of phillyrin as index, Partial Least Squares regression(PLSR) was adopted at original absorption spectra, first derivative spectra, second derivative spectra and WT spectra. Correlation coefficient, RMSEC and RMSEP were regarded as 3 indexes to evaluate modeling. With HPLC as a reference method, PLS was utilized by original absorption spectra, first derivative spectra, second derivative spectra, WT spectra to build the quantitive calibration equation, correlative coefficient of modeling were 0.945, 0.983, 0.960 and 0.977, RMSEP were 1.322, 0.925, 1.126 and 0.914, RMSEC were 0.530, 0.379, 0.481 and 0.390. The results show model established by WT and first derivative spectra were near, the principal component by first derivative spectra was 3, but WT spectra compressed the original spectra data effectively, removed the interference of noise and background, the principle component was 2. Consequently the modeling speed was improved and precision of prediction was increased. And this paper provides an fast and effective data processing method for analyzing the conventional chemical constituents in Chinese medicine.As the research above shown, the NIR technique can detect samples without complex pretreatment rapidly and accurately. The technique owns many advantages that the traditional method don't have. The proposed method can be applied to fast detection and quality control in industrial process. In a word, the application of NIRS technique has good potential in the modernization of Tradition Chinese Medicine.