基于拟分布函数的中药色谱指纹图谱的相似性分析  

Quasi-Distribution-Based Similarity Analysis of Chromatographic Fingerprint of Herbal Medicine

作　　者：安金兵[1] 许玉琼[2] 贺东奇[1] 蔡少青[2] 

机构地区：[1]北京大学医学部公共教学部,北京100083 [2]北京大学药学院天然药物学系,北京100083

出　　处：《北京大学学报（自然科学版）》2008年第5期695-699,共5页Acta Scientiarum Naturalium Universitatis Pekinensis

摘　　要：针对中药色谱指纹图谱中色谱峰面积值存在较大变异性的问题,从拟分布函数的角度介绍了一种新的分析方法。采用秩变换将色谱峰面积值转化为相对秩次,然后通过一系列变换将原数据转换成拟分布函数的形式,再利用柯尔莫哥洛夫检验实现不同类别样品的色谱指纹图谱之间的相似性分析。通过对北马兜铃和南马兜铃样品的比较,说明了该方法有较好的灵敏度和稳健性,因而是比较合理的。Focusing on the problem of high variability of peak area in chromatographic fingerprint of herbal medicine, a new approach on the similarity analysis based on the quasi-distribution function is proposed. Each peak area of chromatographic fingerprint with high variability is transformed into comparative rank. Then after a series of data transformations on original data, the quasi-distribution function is obtained and Kolmogorov test is used to evaluate the similarity of fingerprints of two classes of samples. In the end, by comparison among the samples of Aristolochia contorta and Aristolochia debilis, this method with high robustness and sensibility proved to be reasonable.

关 键 词：色谱指纹图谱 秩 拟分布函数 柯尔莫哥洛夫检验 

分 类 号：R286[医药卫生—中药学]