论文部分内容阅读
目的:运用R型因子与神经网络两种聚类分析方法对收集到得不同产地来源的赤芍药材样品HPLC指纹图谱进行聚类分析,建立对赤芍进行质量判别的分类模型。方法:收集不同产地来源的赤芍样品40个,用HPLC法测定芍药苷含量,结合传统的经验对赤芍进行分类并获得HPLC图谱,运用获得R型因子与神经网络两种方法对不同产地赤芍进行指纹图谱的聚类分析,形成判别函数。结果:两种分析方法结果比较仅19,27,34号样品判别结果有所差异,相似度达91.4%。结论:两种聚类分类方法互相验证,所建立的模型能够为药材的质量评价提供一个快捷、准确、可行的鉴别方法。
OBJECTIVE: To establish a classification model for quality identification of Radix Paeoniae Rubra by using two kinds of cluster analysis methods of R-factor and neural network to cluster the HPLC fingerprints of Radix Paeoniae Rubra from different origins. Methods: Forty samples of Paeoniae Radix were collected from different origins. The content of paeoniflorin was determined by HPLC. The traditional Chinese medicine was used to classify Radix Paeoniae Rubra and obtain the HPLC profile. By using R-factor and neural network, Shao fingerprint analysis of the cluster analysis, the formation of discriminant function. Results: Comparison of the results of the two methods showed that the discriminant results of samples No. 19, No. 27 and No. 34 were all different, with a similarity of 91.4%. Conclusion: The two clustering methods verify each other. The established model can provide a quick, accurate and feasible identification method for the quality evaluation of medicinal materials.