高效液相色谱法结合支持向量机鉴别桑白皮

Identification of Sang-Bai-Pi(Cortex Mori Albae Radicis) by combined high performance liquid chromatography and Support Vector Machine Method

目的建立一种区分基原为Morus alba L.和非M.alba L.的中药桑白皮的可靠鉴别方法。方法以桑白皮乙醚部位高效液相色谱(high performance liquid chromatography,HPLC)法色谱峰的归一化数据作为输入向量,随机分为训练样本和测试样本,用支持向量机软件构造区分不同基原样本数据的分类超平面以建立分类模型,并将分类结果与相似度分析法相比较。结果相似度分析法未能对某些桑白皮样本是否属于M.alba L.给出准确判断。支持向量机方法所建分类模型,对样本属于M.alba L.与否的预测准确度达到100%。结论高效液相色谱法结合支持向量机能够准确鉴别桑白皮的基原属性,在中药鉴别领域有广阔的应用空间。

Objective To develop a reliable method for recognizing Cortex Mori Radicis （CMR ） samples originated from Morus alba L. or not. Methods Using normalized high performance liquid chromatography（HPLC） peak areas of CMR ether extracts as modeling data, and using two randomly divided sample-sets as training data and testing data, the optimal separating hyperplane of two datasets and the classification model were constructed by support vector machine（SVM） software. For comparative purpose, the similarity of HPLC profiles was calculated using vectorial angle method. Results Confirmation of M. alba origin of above CMR samples was partly missed in similarity analysis, while the predicted classification accuracy of those samples reached 100% when using SVM classfication model. Conclusion Combined HPLC and SVM method can correctly recognize the original plant features of Sang-Bai-Pi,and can be easily generalized to identify more other Chinese medicinal materials.