20种活血化瘀中药提取物的红外光谱分析及其与药性相关性的初步探讨  

Analysis of the Properties and Relationships Between the Infrared Spectra of 20 Kinds of TCM

作　　者：邓家刚[1] 刘进[2] 覃洁萍[1] 冯旭[1] 王丽丽[1] 侯小涛[1] 

机构地区：[1]广西中医学院,南宁市明秀东路179号530001 [2]广西师范学院,南宁市燕子岭路4号530023

出　　处：《光谱实验室》2010年第2期741-744,共4页Chinese Journal of Spectroscopy Laboratory

基　　金：国家973计划资助项目(2007512602)

摘　　要：为探讨中药提取物红外光谱用于中药药性识别的可行性,分别采用石油醚、乙酸乙酯、无水乙醇、水4种溶剂分部位提取桃仁等20种不同药性活血化瘀中药的化学成分,并用FTIR技术对各溶剂提取得到的成分进行分析测定,共采集得到240个傅里叶变换红外图谱。经对原始数据预处理后,应用主成分分析结合支持向量机方法对各样本的IR图谱进行分析。结果显示用乙酸乙酯提取物的红外光谱构建的模型对平性药识别正确率为80%,对非平性药的识别正确率为90%,平均识别正确率为85%。In order to study the feasibility of application of IR spectra to traditional Chinese medicine （TCM） property recognition, four different solvent sequences （petroleum ether, ethyl acetate, ethanol and water ） were selected to extract the chemical components in 20 kinds of huoxuehuayu-TCM. After removing the solvents, the extracts of different solvent portion were determined by Fourier transform infrared （FTIR） spectrometry,and 240 IR spectra were obtained. Subsequently, the raw data were preprocessed, and the relationships of the properties of TCM and infrared spectra were studied by support vector machine （SVM） and principal component analysis （PCA） methods. The results showed the prediction model established by the IR spectra of ethyl acetate portion have the higher recognition accuracy, the recognition accuracy for the neutral medicine in prediction set was 80%, and 90% for non-neutral medicine, which made up accuracy of 85% in average.

关 键 词：主成分分析 支持向量机 红外光谱 中药药性识别 

分 类 号：O657.33[理学—分析化学]