| Rhubarb is a traditional medicine having many medicinal effects, such as antibiotic, protecting liver, and so on. In order to ensure that unofficial rhubarb samples are not mixed into official rhubarb in commercial products, and active constitutes in rhubarb are determined exactly, quality control of rhubarb is very necessary. The traditional identification of medicinal material is based on shape, configuration and microscopic feature. The analysis of other methods needs separation and purification, which consumes time and reagent. So it is need to develop a quick, easy, modified method. The fast and non-destructive near-infrared spectra (NIRS) are suitable for quality control of Chinese medicinal materials. But NIRS is an indirect method of measurement, and it needs analysis by many chemometrics methods.In the present paper, chemometrics methods applied in rhubarb NIRS and solid substrate room temperature phosphorescence (SS-RTP) are studied. Multivariate statistical analysis such as principle component analysis (PCA) and SIMPLe-to-use interactive selfmodeling mixture analysis (SIMPLISMA) and so on, and radial basis function (RBF) neural network, ridge regression RBF (RRRBF) neural network, support vector machine (SVM) are applied. Traditional identification and new technology are combined to set up new calibration methods, by which the analysis of intricate system including Chinese medicine, and organic compounds can be realized. The results as follows:1. Qualitative and quantitative analysis of rhubarb by NIRSIn qualitative analysis, NIRS of rhubarb are studied by principle component analysis (PCA). The results of studies show that the PCA of rhubarb samples supports a reliable science data. The NIRS are processed by second-derivative to enhance spectral resolution and eliminate slop background. And the NIRS are also wavelet-compressed to reduce variables and thus to speed up computation. The... |
PDF Full Text Request