基于潜在语义分析的中药分类技术  

Classification Technology of Traditional Chinese Medicine Based on Latent Semantic Analysis

作　　者：龙长江[1] 黎煊[1] 刘浩鹏[1] 

机构地区：[1]华中农业大学工学院,湖北武汉430070

出　　处：《中国中医药信息杂志》2015年第3期5-9,共5页Chinese Journal of Information on Traditional Chinese Medicine

基　　金：国家自然科学基金(61007058)

摘　　要：中医理论现代化是中医发展过程中的关键问题,但中医概念过于模糊,缺乏精确定义,以致难以进行技术验证。迄今,中药分类方法不为西方医学所承认,鉴于中药成分的多样性和复杂性,难以用生物化学方法逐一对中药分子成分进行分离、药效验证并分类。为验证中药分类的科学性,将药物的分子特性转换到光谱中进行研究,结合互联网搜索引擎采用的潜在语义分析技术结合加权方法,对滋阴、壮阳、补气和补血4类中药共48种药材的近红外光谱数据进行研究,分类结果与中药分类近似。Modernization of theories of traditional Chinese medicine （TCM） is a key issue for the development of Chinese medicine. As TCM concepts are too vague and lack precise definition, it is difficult to test those theories by modern technologies. Methods of TCM classification are not accepted by Western medicine so far. Because of the diversity and complexity of Chinese herbal medicines, the molecules of Chinese herbal medicines are difficult to be separated, and efficacy is difficult to be verified and classified by biological and chemical methods. In order to verify the scientificity of TCM classification, molecular properties can be studied by their spectrum, and be combined with latent semantic analysis technology and weighting method which are widely used in the Internet search engine. 48 Chinese herbal medicines belonging to four kinds of Chinese herbal medicines：nourishing yin, tonifying yang, invigorating qi, and enriching blood, are analyzed through NIRS, and the results are almost the same as the TCM classification.

关 键 词：中药 潜在语义 近红外光谱 

分 类 号：R2-05[医药卫生—中医学]