60种中药脂类成分GC-MS图谱与寒热药性关系的统计识别模式研究  

Study on Statistical Pattern Recognition Model for Relationship Between Cold-Heat Natures and Lipid Based on GC- MS of 60 Traditional Chinese Medicines

作　　者：王晓燕[1] 李峰[2] 

机构地区：[1]山东中医药大学理工学院,山东济南250355 [2]山东中医药大学药学院,山东济南250355

出　　处：《辽宁中医杂志》2015年第6期1303-1305,共3页Liaoning Journal of Traditional Chinese Medicine

基　　金：国家重点基础研究发展计划(973计划)资助项目(2007CB512601);山东中医药大学2013年"名科工程"青年骨干培养计划(ZYDXY1301)

摘　　要：目的:探讨中药寒热药性与所含脂类相关性。方法:取寒性、热性中药各30种,提取粗脂并按1 m L/0.1g加入2%硫酸-甲醇在70℃下甲酯化,GC-MS联用测定脂类图谱,并利用5种判别分析方法进行药性判别建模。结果:SVM方法优于其它判别方法,组内判别一致率为100%,5折交叉验证正确率53.2%,测试集预测率58.3%。结论:中药寒热药性与脂类成分存在一定相关性,脂类是中药药性的物质基础之一;SVM可以作为判别中药寒热药性的辅助工具。Objective ： To discuss the correlation between cold-heat natures and lipid. Methods ： We extracted lipid from 60 tra- ditional Chinese medicines （TCM）. The lipid was determined by pulsing 2% sulfuric acid -methanol as 1 mL/0. 1 g, and methyl etherification under 70℃ water bath. After that, the study data was determined by GC - MS method. Finally the results were ana- lyzed by 5 well-known classification methods. Results ： Statistical simulation showed that SVM model had higher sensitivity than the other models. In SVM model, discrimination consistent rate in the group was 100% ,5 fold cross -validation accuracy 53.2%, prediction rate was 58.3% in test set. Conclusion：The SVM could distinguish the cold and heat natures of TCM. There is a certain correlation between cold - heat natures and lipid. In addition, the lipid is one of the material base of natures of TCM.

关 键 词：脂类 GC—MS图谱 寒热药性 SVM 

分 类 号：R285.1[医药卫生—中药学]