Weng Xin-Xin, Zhang Zhong-Hu, Yin Li-Hui, Lu Feng
School of Pharmacy, Second Military Medical University, Shanghai 200433, China.
Guang Pu Xue Yu Guang Pu Fen Xi. 2010 Apr;30(4):984-7.
In the present paper, five different kinds of hypoglycemic tablets were identified using kernel principal component analysis (KPCA)-clustering analysis of their Raman spectra. KPCA was used to compress thousands of spectral data into several variables and to describe the body of the spectra before clustering analysis was chosen as further research method. The results showed that hypoglycemic tablets could be quickly classified using KPCA-clustering analysis. A disadvantage of Raman spectroscopy for this type of analysis is that it is primarily a surface technique. As a consequence, the spectra of the tablet core and its coating might differ. However, the KPCA-clustering analysis turned out to be a sufficiently reliable discrimination, i. e., 96% of the hypoglycemic tablets with coating and 100% of the hypoglycemic tablets without coating were predicted correctly. Overall, the Raman spectroscopic method in the present paper plays a good role in the identification and offers a new approach to the rapid discrimination of different kinds of hypoglycemic tablets.
在本文中，利用核主成分分析（KPCA）-拉曼光谱聚类分析对五种不同的降糖片进行了鉴别。KPCA用于将数千个光谱数据压缩为几个变量，并在选择聚类分析作为进一步研究方法之前描述光谱主体。结果表明，采用KPCA-聚类分析可快速对降糖片进行分类。拉曼光谱用于此类分析的一个缺点是它主要是一种表面技术。因此，片剂核心及其包衣的光谱可能会有所不同。然而，KPCA-聚类分析结果显示出足够可靠的鉴别能力，即带包衣的降糖片96%以及无包衣的降糖片100%被正确预测。总体而言，本文中的拉曼光谱方法在鉴别中发挥了良好作用，并为快速鉴别不同种类的降糖片提供了一种新方法。
