论文部分内容阅读
Objective: As a complementary medical system to Western medicine, traditional Chinese medicine (TCM) provides a unique theoretical and practical approach to the treatment of diseases over thousands of years.Confronted with the increasing popularity of TCNM and the huge volume of TCM data, historically accumulated and recently obtained, there is an urgent need to explore these resources effectively by the techniques of knowledge discovery in database (KDD).This paper aims at providing an overview of recent KDD studies in TCNM field.Methods: A literature search was conducted in both English and Chinese publications,and major studies of know(edge discovery in TCM (KDTCM) reported in these materials were identified.Based on an introduction to the state of the art of TCM data resources, a review of four subfields of KDTCM research was presented, including KDD for the research of Chinese medical formula, KDD for the research of Chinese herbal medicine, KDD for TCM syndrome research, and KDD for TCM clinical diagnosis.Furthermore, the current state and main problems in each subfield were summarized based on a discussion of existing studies, and future directions for each subfield were also proposed accordingly.Results: A series of KDD methods are used in existing KDTCM researches, ranging from conventional frequent itemset mining to state of the art latent structure model.Considerable interesting discoveries are obtained by these methods, such as novel TCM paired drugs discovered by frequent itemset analysis, functional community of related genes discovered under syndrome perspective by text mining, the high proportion of toxic plants in the botanical family Ranunculaceae disclosed by statistical analysis, the association between M-cholinoceptor blocking drug and Solanaceae revealed by association rule mining, etc.It is particularly inspiring to see some studies connecting TCM with biomedicine, which provide a novel top-down view for functional genomics research.However, further developments of KDD methods are still expected to better adapt to the features of TCM.Conclusions: Existing studies demonstrate that KDTCM is effective in obtaining medical discoveries.However, much more work needs to be done in order to discover real diamonds from TCM domain.The usage and development of KDTCM in the future will substantially contribute to the TCM community, as well as modern life science.