Study Of Classification-Association Rules On Zhongjing Formulas' Data Mining

Zhang Zhongjing wrote "Shang Han Za Bing Lun" , which was divided into "Shang Han Lun" and "Jin Kui Yao Lue" . These two books founded from the view point of theory and practice the pioneering work of syndrome differentiation of the six meridians, established a systematic diagnosis and treatment methodology combining syndrome differentiation and disease identification of Traditional Chinese Medicine. The formulas were summarized carefully and proven effective. It is thereby considered as the classic in prescription. Data mining is a newly arisen discipline, later which emerged recently along with the maturity of database technique and the new development of computer storage deficient. It can solve the antinomy of present society——data abundant but information needy. So to make use of data mining technique to study Traditional Chinese Medicine formulas, especially to explore the classical prescriptions, is an endeavor with very great expectation and intense challenges. In the meantime it will provide a propelling force to prove and elaborate academic thoughts of Zhang Zhongjing, to promote the research of basic theory and methodology of Traditional Chinese Medicine, to raise the academic research level of whole Traditional Chinese Medicine, and to expand space the existence of the Traditional Chinese Medicine.In this study we chose the two books' formulas as the data source firstly. Then we built a database of Zhongjing formulas to include the information of ID, name, source, effect, therapeutics and the medicine constitute. Secondly, we classify these data according to the principle in the basic theory of Traditional Chinese Medicine and characteristics of Zhongjing formulas to correspond to the nature of disease, the location of disease, the symptoms of disease and the source of the formulas. Thirdly, we initiate for the first time to apply data mining technique to exact information, especially the data preprocessing technique including data cleaning, data integration, data transformation and data reduction. As result, we eliminated the vacant and repeated data and the phenomena of the same medicine with different names and different medicine with the same name. The data was standardized that to meet the need of subsequent mining. Finally, we make use of the technique in extracting classification-association rule to analyze the standardized data to obtain knowledge concealed inside them. Result: ZhangZhongjing chose the herbal drugs according to the nature of diseases, the location of diseases and the symptoms of diseases for the cure of diseases. Especially, some drug's compatibility reflects his special viewpoints in treating specific diseases. For example, ZhangZhongjing chose medicine such as Schisandra chinensis, asarum, cassia twigs, shrubby to cure the cold syndrome. He chose medicine such as shrubby, apricot kernel, liquorice together more frequently to cure the exterior syndrome. He chose medicine such as cassia twigs, ginger, liquorice, jujube, rhubarb together more often to cure the taiyang syndrome. He chose medicines such as schisandra chinensis, asarum, cassia twigs, shrubby together more often to cure the disease caused by exogenous pathogenic factors.In summary, we used successfully data preprocessing technique and classification-association rule to analyze the formulas of ZhangZhongjing. It is made clear that data mining can be applied to study classic prescriptions and to find knowledge that concealed inside them. Not only the knowledge can enrich the basic theory of Traditional Chinese Medicine but also guide the clinic of practice of Traditional Chinese Medicine. Therefore it can accelerate the development of Traditional Chinese Medicine. At the same time we arrived at a conclusion that data mining technique is a kind of advanced information processing technique, which can help people to discover certain modes and regulations in Traditional Chinese Medicine, especially in the Traditional Chinese Medicine formulas.