Study On The Identification And Analysis Of The Combination Of The Instrumental Analysis With The Chemometrics

Since being invented, the chemometric developed quickly and a lot of great methods in recognition analysis were found and developed. With the increasing complexity of analytical data, traditional data processing methods for spectral analysis can not meet the analysis need. The chemometrice uses methods of mathematics, statistics and computer science to design optimum chemical measurement procedures and to provide maximum information about complexity of analytical data. In the present papers introduce the history of chemometric and the principle and the application of several commonly used method in recognition analysis(SVM, PLS, PCA, ANN, CA). At the same time, in this part also makes many review about worldwide research results of the salt, pueraria and bletilla, it can provide reference for study as well.The quality of the salt attention since it is become the most popular thing day by day in our daily life, however, the method of quality identification still re quires further study. A fast, effective and low-cost analysis method using NIR spectroscopy with PLS and SVM of the chemometric for quality identification of five different kinds of the salt was proposed in the second chapter. By comparing two different kinds of calculation processes, finally find the best calculation process for the quality identification of the salt.Pueraria and bletilla as the widely known Chinese medicine for a lot of people, meanwhile, not only the quality of the pueraria and bletilla but also its nutrients are attracted more and more people attention. But, the method of quality identification still need further study. Fluorescent and IR was considered to identifying the different kinds of pueraria and bletilla since these high sensitivity, simple operation and good stability. The pueraria was detected by the IR spectroscopy and three-dimensional fluorescent in chapter three, then the analysis of the bletilla was in the chapter four. Single factor experiments to select the appropriate condition of the extract process. After this, 5 kinds of pueraria and bletilla were extracted under this best condition and then scan them in IR spectroscopy and three-dimensional fluorescent spectra. PLS and PCA methods used to build the model to identify the obtained data. The accuracy of the model is well, which proved this is a viable method for detection the kind of pueraria and bletilla.At the same time, the pueraria and bletilla is the most serious fraud in the market, it is also the most difficult to detect. There is few reports or research of concern the issue of how to detect the purity in pueraria and bletilla. Chapter three and chapter four discusses using IR spectroscopy and three-dimensional fluorescent spectroscopy to class adulterated pueraria and bletilla. The classification models for the five kinds of the adulterated pueraria and bletilla were built by PCA and PLS, respectively. The accuracy of the model is satisfactory and the results ensure that IR spectroscopy and three-dimensional fluorescent spectroscopy combined with PCA and PLS can be utilized for determination of adulteration in pueraria and bletilla.