模糊聚类分析法在郁金裂解色谱测定指纹图谱中的应用研究

Application of fuzzy cluster analysis in determining fingerprint spectrum of curcuma aromatica salib by pyrolysis gas chromatography

采用模糊聚类分析法对10个郁金样品的裂解色谱图在16个峰参数上进行分类,并讨论了6种聚类方法的分类效果。分类所得结果与实际相符。研究表明该方法能全面、综合地反映图谱间的相似性关系,从而为中药指纹图谱的建立提供了有力的依据,也为准确地选出有代表性的图谱提供了一种直观、便捷的方法。

Fuzzy cluster analysis is applied to classing 10 samples of Pyrolysis gas chromatography of Curcuma aromatica Salisb with 16 peak parameters. The effect of six kinds of cluster methods is discussed in this paper. The cluster result matches, which indicates that this method can reflect the resemblance relationship between spectrums well. Thereby, it provides convincing evidence for the establishment of fingerprint. Moreover, it offers an intuitionistic, quick and convenient way for choosing typical spectrums correctly.