Identification of the mineral oil fluorescence spectroscopy based on the PCA and ICA-SVM

Abstract

Three-dimensional fluorescence spectroscopy technology is often used to identify the kind of the mineral oil. The dimension of it is high which cause the characteristic of the oil style-book are difficult to be maintained by the simple formula. In this paper, the principal component analysis (PCA) is used to reduce the dimensions of the spectroscopy. The independent component analysis (ICA) is used to do the matrix decomposition from the perspective of independence to extract the main feature of the spectroscopy data processed by the PCA. The support vector machine (SVM) is used to assort the main characteristic root books which are abstracted by the ICA. The species identification of the mineral oil will be realized by it. The identification result is visualized by the parallel coordinate's graph. The experiment results show that it is effective to extract the main feature of the spectroscopy. The classify speed is greatly increased. The identification of the oils can be realized with high discrimination which is 99.12%.