中文摘要：

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。

英文摘要：

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.