作者: Zhihong Xu , Yan Liu , Xiaoyong Li , Wensheng Cai , Xueguang Shao
DOI: 10.1016/J.SAA.2015.05.030
关键词:
摘要: Abstract Principal component discriminant transformation was applied for discrimination of different Chinese patent medicines based on near-infrared (NIR) spectroscopy. In the method, an optimal set orthogonal vectors, which highlight differences between NIR spectra classes, is designed by maximizing Fisher’s function. Therefore, a model discriminating class and others can be obtained with tiny classes. Furthermore, because contain large amount redundant information, principal analysis (PCA) employed to reduce dimension. On other hand, continuous wavelet transform (CWT) taken as pretreatment method remove variant background. For identifying same medicine from manufactures were studied. The results show that all models provide 100% discrimination.