基于随机森林的中药寒、热药性代谢组学判别方法研究  

The Research for Metabolomics Discriminant Method for Cold and Hot Property of Traditional Chinese Medicine Based on Random Forest

作　　者：聂斌[1] 郝竹林[1] 桂宝 王卓[3] 杜建强[1] 王国龙[1] 张鑫[1] 

机构地区：[1]江西中医药大学计算机学院,南昌330004 [2]江西省公安厅科技信息化总队,南昌330006 [3]南昌大学软件学院,南昌330047

出　　处：《江西中医药大学学报》2015年第2期82-86,共5页Journal of Jiangxi University of Chinese Medicine

基　　金：国家自然科学基金项目(61363042);江西省教育厅科技计划项目(GJJ13014);江西省卫生厅中医药科研计划项目(2013A023;2013A065);江西中医药大学校级课题(2013ZR0068)

摘　　要：目的:建立中药寒、热药性判别模型与方法。方法:利用中药寒、热药动物实验,获取代谢组学数据;再采用随机森林算法构建中药寒、热药性分类判别模型。结果:基于随机森林构建的中药寒、热药性代谢组学分类判别模型,能够很好地实现分类判别,总体准确率超过90%;用前30个最重要的M/Z值构建的分类判别模型,同样有很高的分类准确率;经7∶3测试,准确率也超过90%。结论:基于随机森林的中药寒、热药性代谢组学分类判别模型,经实验数据建模验证表明其可行有效。Objective： To establish discriminant method for cold and hot property of traditional Chinese medicine.Mthod：Obtained metabolomics data from experiments on animals.Built model for discriminant cold or hot property of traditional Chinese medicine based on random forest.Results：The model could well realized itsˊclassification, the overall accuracy was more than 90%.The same high level accuracy use the top 30 M/Z values to build model.The same high level accuracy used 7∶3 test.Conlusion： The metabolomics discriminant method for cold or hot property of traditional Chinese medicine based on random forest, which is proved to be feasible and effective after tested with the experiments data.

关 键 词：随机森林 中药药性 代谢组学 分类 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]