雷达图表示与相似度计算相结合的麻黄品种鉴别方法

Identification of ephedra species by radar graph representation and similarity calculation

将指纹图谱技术用于中药麻黄的鉴别,提出了指纹图谱雷达图表示与相似度计算相结合的麻黄品种鉴别分析方法。该方法首先对麻黄样品的HPLC指纹图谱的峰面积分别进行相似度计算和雷达图表示,然后对雷达图进行重心图特征提取,最终得到用于分类比较的雷达图特征,最后结合雷达图特征与相似度特征对麻黄品种进行鉴别。实验结果显示该方法鉴别准确率明显高于只采用相似度模式特征的鉴别,且综合特征的最高分类准确率达98.25%,表明雷达图表示与相似度特征相结合的方法是一种可行的麻黄品种鉴别方法。

The fingerprint technique was applied to identification of ephedra species, and a novel ephedra species identification method based on fingerprints＇ radar graph representation and similarity calculation was advanced. The method is characterized by the following. Firstly, the peak areas of ephedra HPLC fingerprints are analyzed by similarity calculation and radar graph. Secondly, the radar graph featuers are obtained for classification and comparison. Finally, the radar graph features and similarity features are combined to identify the ephedra species. The accuracy of this combining method is better than the identifications using similarity features, and its highest accurate classification rate is 98.25%. Thus, the method combining radar graph representation with similarity features is a feasible ephedra species identification method.