The Application Of Chemometrics In Analyzing Chemical Information

Wavelet analysis was a novel and powerful mathematical tool which has been developed sharply in the eighties of the twentieth century. With the good localization and flexible time-frequency windows, it was regarded as the crystal work of pure mathematical field and a milestone in the developing history of Fourier transform. At the same time, with the time-frequency localization, scale change and direction, it was also wildly applied in the fields of applied mathematics and technical science as the great breakthrough of tools and methods in computer identification, pattern recognition, signal analysis, signal-picture analysis, quantum physics, radar, earthquake wave analyzing, celestial mechanics, edge value detection, nonlinear science and engineering technology in recent years. Wavelet analysis, which evolved from Fourier transform and actually extended and developed it, embodied another perfect combination between the pure mathematical and the applied mathematical behind the Fourier analysis.The application of wavelet analysis has penetrated in all fields of science with the mature of its theories little by little and the continuous perfection of corresponding software. In analytical chemistry, its application mainly focused on some areas as follows: (1) data smoothing and denoising; (2) solve the overlapping signal, rectify baseline for spectra, detect the edge value; (3) image analysis and compression; (4) extract the valuable information of signal by combining with some other analytical technology such as the artificial neural network. All of these injected the new energy into the analytical methods.The artificial neural network, as an information analytical system by imitating the configuration and action of human neural network, was proposed by the studied result about the human brain configuration, and essentially different with the traditional information analytical technology. Based on the theory of mathematicaltopology, its mainly characteristic such as dealing with large numbers of datum, higher ability of containing error, processing and saving the information, the function of self-organized and self-trainning imitated the information processing course of human brain, its velocity and ability of dealing with the complex information further exceeded traditional methods. At present, its application acquired conspicuous achievement in some fields such as signal analysis, pattern recognition, target trace, robot control, expert system, combination optimize, network management, so on and so forth. In the areas of chemistry, biological technique and medicine study, its application also development sharply.The fractal theory which aimed at the non-le;vel and non-regular geometry in nature, was an very active embranchment in non-linear science. It admitted that some local or process resembled with its whole at some conditions at the shape, configuration, function, information, time, energy, and so on, the change of its dimension might be discrete integer or continuous decimal fraction, this enlarged the view of study. From the seventies of the twentieth century to now, its application has involved in almost all of the fields of the nature science and society science, became a prior study course in many international problem.The main works of this paper were as follows: Part I: introductionThis part narrated the developing history of Wavelet analysis, artificial neural network and fractal theory, introduced their basic mathematical theory and summed up their application in the field of analytical chemistry. Here, the basic theory of Wavelet analysis for detecting the catastrophe information and analyzing the image, the training arithmetic of self-organized feature map, and the calculating method of fractal dimension were showed. One hundred and one references were cited in this part.Part II: Wavelet analysis was applied for analyzing the frequency information of complex systemThe oscillating frequency spectrum, which generated from the continuous wavelet transform, was applied to analyze the frequency information...