Application Of Wavelet Analysis And Chemometrics Methods In Analyzing Traditional Chinese Medicine

Multivariate calibrations and multivariate curve resolutions, the main body of chemometrics are reviewed at the first part of this dissertation. A brief review on the applications of chemometrics in the Traditional Chinese Medicine is also given. In this dissertation a series of novel chemometrics algorithms, combined of newly developed wavelet analysis and original chemometrics algorithms are developed and be applied in the complex chemical analytical systems as TCM (Traditional Chinese Medicine ). These algorithms could be widely used for different analytical instruments or different analytical objects and it covered from ultra-violet data to complex HPLC-DAD chromatographic data.A lot of research work have been done on some branches of chemometrics, such as multivariate calibrations, multivariate curve resolutions and chemical pattern recognitions. Wavelet analysis is introduced to chemometrics. The main aims of the research are to measure the effective components in the complex TCM systems more precisely with modern analytical instrument and at the lower requirement of separation level. The main contents include the following:1. Multivariate Calibrations: The denoising abilities of wavelet transform are studied in the multivariate calibrations. Some novel chemometrics algorithms are provided, such as multi-resolution analysis, thresholding method based on wavelet transform, max mode wavelet transform and wavelet packet transform coupled with principal components analysis. With the combination of non-classic optimization methods, these algorithms can give better results. For example, the artificial neural net (self organized map ANN) has been used in the searching of optimizing wavelength and best set of wavelet coefficient set, and genetic algorithm has been used in the wavelength selection, searching for best wavelet bases and optimization of wavelet coefficients.2. Multivariate Curve Resolutions: Detail comparison of different criterions of local factor analysis in the multivariate curve resolutions is provided. The wavelet analysis technology is also introduced to moving window factor analysis. Wavelet moving windows factor analysis and wavelet target windows factor analysis are derived. Some new methods for peak purityassessment are introduced, and compared with some classic methods.3. Chemical Pattern Recognitions: Wavelet transform is introduced to chemical pattern recognitions. With successful combination of the wavelet multi-resolution ability and fractal theory, the TCM injections are classified satisfactorily. The method is expected to be used in the quality control of TCM and classification of products from different companies.