Application Of Chemometrics In Recognition Analysis

Since being invented, chemometric developed quickly and many greatmethods in recognition analysis were developed. Chapter one introduce thehistory of chemometrics and talk about the principle and application of severalcommonly used method in recognition (SVM, ANN, PCA, CA). Also make areview about worldwide research results of tea, honey and acetonitrile, it canprovide reference for study.The quality of tea and honey attentions since they are two most popularthings for drink, but the method of quality identification requires further study.A fast, effective and low-cost method using NIR spectroscopy with SVM forquality identification of tea and honey was proposed in chapter two. Bycomparing three different calculation processes, finally find the best calculationprocess.Honey is the most serious fraud in the market, it is also the most difficultto detect. Fluorescent was considered for identifying the kind of honey since itshigh sensitivity. But the fluorescence generated from the honey is very weak, itneeds to be derived before detect. The honey was derived with ethanolamine,detected by the three-dimensional fluorescence in chapter three. Single factorexperiments to select the appropriate condition of derivative reaction. After this,4kinds of honey were derived under this condition and then scan them inthree-dimensional fluorescence spectra. SVM method used to build the model toclass the honey using the obtained data. The accuracy of the model is good,which proved this is a viable method for detection the kind of honey.Acetonitrile is one of the most commonly organic solvent, even though thehigh purity acetonitrile require in many ways, but few reports of concern the issue of how to detect the purity in acetonitrile. Chapter four discusses usingnear-infrared spectroscopy to detect five kinds of impurities incorporated inacetonitrile. The classification models of Kind of acetonitrile, Kind of impurityand the amount of impurity were built by SVM, respectively. The accuracy ofthe model is satisfactory and the results ensure that NIR spectroscopy combinedwith SVM can be utilized for determination of adulteration in acetonitrile.