摘要
目的:运用电子鼻提取中药气味特征来建立气味指纹图谱,同时建立气味指纹图谱的鉴别模型,实现对中药的快速鉴别,提供一种中药气味鉴别新方法。方法:以传感器阵列的最高响应值为指标,运用逐步判别法对气味指纹图谱进行特征提取并建立鉴别模型,并将其运用于对姜科白豆蔻、草豆蔻、草果、莪术、干姜、高良姜、姜黄、砂仁、益智和郁金等10味中药的气味鉴别。结果:实验数据显示逐步判别法建立的鉴别模型对训练集中100个样本和测试集中20个样本的回代正判率分别为97%和90%,鉴别效果满意。结论:应用逐步判别法建立气味指纹图谱的鉴别模型准确、有效、可行,能很好地鉴别不同品种的中药。该方法为现行中药质量标准中气味指标的数字化、客观化提供实验依据和方法;也为中药指纹图谱的数据挖掘提供新的研究思路。
Objective： To establish a fingerprinting of scent in traditional Chinese medicine （TCM） based on electronic nose. To build up a new identification model of the scent fingerprinting and provide a new method for the identification of TCM. Method ： The peak value of sensor array was chosen as the statistical index while stepwise discriminant analysis, a classic modeling method for chemical pattern recognition, was employed to extract the characteristic parameters from the scent fingerprinting and built up an optimized identification model, which was applied to distinguish different species of TCM. Results ： It is illustrated that the correct judge rate of train set with 100 samples and test set with 20 samples was 97% and 90%, respectively. Conclusion： In conclusion, the established identification model in this research was convenient, efficient and feasible to identify a variety of TCM. Experimental data and method were provided for the standardization of TCM＇s scent in the current quality standard. This research may give insight to offer a guideline for data mining in TCM fingerprinting using any other detection means.
出处
《中国药品标准》
CAS
2012年第5期340-343,共4页
Drug Standards of China
基金
北京中医药大学自主选题项目(JYB22-XS041)
关键词
逐步判别法
中药
气味指纹图谱
模式识别
鉴别模型
stepwise discriminant analysis
traditional Chinese medicine
scent fingerprinting
pattern recognition
identification model
