中药组方的计算机辅助分类与识别  

Computer-aided Classification and Identification of Traditional Chinese Medicine Herbal Formula

作　　者：蔡从中[1] 袁前飞[1] 肖汉光[1] 刘兴华[1] 孔春阳[2] 

机构地区：[1]重庆大学数理学院 [2]重庆师范大学物理学与信息技术学院,重庆400047

出　　处：《重庆大学学报（自然科学版）》2006年第10期42-46,共5页Journal of Chongqing University

基　　金：重庆大学与新加坡国立大学国际联合科研资助项目(ARF-151-000-014-112);重庆市自然科学基金项目资助(CSTC;2006BB5240);重庆大学基础及应用基础研究基金资助项目(71341103)

摘　　要：应用支持向量机(SVM)在中药药方的识别和新中药药方的预测方面进行了有益的探索.介绍了支持向量机的应用领域,以及多味传统中药组方的数据样本的收集、草药的特征向量构造方法,给出了SVM★对4味-10味复方的训练及测试结果等.被预测为假阳性(False positive)的样本极有可能是迄今为止人们尚未发现的新药药方,因此将具有提供给中医药学者进行进一步研究的价值.It is a meaningful investigation on the recognition to a valid Traditional Chinese Medicine herbal formula and the prediction to a new available herbal formula by using support vector machine. After a brief introduction to the application fields of support vector machine, positive samples database collection method of the multi-herbs TCM formulas, and the extraction method for the feature vector of herbs are described in detail. The SVM^＊ training and testing results to the formulas which consist of multi-herbs from number of 4 to 10 are provided and discussed. Finally, it is concluded, to some extent, a false positive formula identified by SVM may be a new potential valid formula, which is still not discovered by humanity and of some value to be further studied by TCM experts.

关 键 词：传统中药 药方 分类 识别 支持向量机 

分 类 号：R2[医药卫生—中医学] R289.1