基于相似系统理论的相似度计算方法的改进  

Improvements in Similarity Algorithms Based on Similarity System Theory

作　　者：詹雪艳[1] 史新元[1] 展晓日[1] 王耘[1] 乔延江[1] 

机构地区：[1]北京中医药大学中药学院,北京100102

出　　处：《分析化学》2010年第2期253-257,共5页Chinese Journal of Analytical Chemistry

基　　金：北京市科技计划课题(No.D02050040040111)资助

摘　　要：针对常用的相似度方法对色谱指纹图谱数据的相对差异响应不灵敏,本文对基于相似系统理论的改良程度相似度的计算方法进行了改进。模拟数据和乳块消片色谱指纹图谱数据计算结果表明,改进后的新改良程度相似度q′可区分与参照样本的总差异相同,但差异分布不同的样本。在共有峰峰面积差异不超过100%的情况下,新改良程度相似度q′计算方法能够更灵敏地反映两样本的峰面积的相对差异。本研究的新改良程度相似度q′能突出组分群中某些组分对既定配比的较大偏离,适用于中药成分群配比波动的控制和中药过程质量控制。For solving the problem that common similarity algorithms are insensitive to the relative differences of different chromatographic fingerprints data,this paper improved the algorithms of improved extent similarity,which named new improved extent similarity.It was confirmed that new improved extent similarity could distinguish the samples with the same total difference and different distribution of relative differences.by applying the similarity algorithm to simulated data and the chromatographic fingerprints of nine batches Rukuaixiao tablets.New improved extent similarity could reflect the relative differences of peak areas more sensitively when the relative differences of common peak areas was less than 100%.The result indicates that new improved extent similarity can reflect the large deviation from the established proportion of components,which can be applied to the control of the fluctuation of the established proportion of effective components and the process quality control of Traditional Chinese Medicine.

关 键 词：相似系统理论 相似度算法 过程质量控制 

分 类 号：R284[医药卫生—中药学]