Data Mining Of A Number Of Ways In Chinese Medicine Database

Traditional-Chinese Medicine has a long history and is based on a huge amount of valuable pharmaceutical information and documents, which is composed of ambiguous words with overloaded details. How to use them to serve for modern pharmacy is a big challenge.The objective of data mining is to discover unknown, valid and usable knowledge from a mass of information. The typical knowledge discoverable by data mining is primarily association rules, classes and clusters. Traditional-Chinese medicine prescriptions are our country's precious medicine mine whose complexity and huge amount of information suit well the application of data mining. The study is aimed at discovering the hidden pharmaceutical principles in the Traditional-Chinese Medicine prescriptions. For this use, the methods of association rules and cluster analysis are presented and applied.In this paper, the characteristics of Traditional-Chinese Medicine prescriptions are studied, and a Traditional-Chinese Medicine prescription database is built. After that, basic principles and methods of Data mining are presented, some data mining algorithms are discussed and applied to the database. In the mining of association rules, Apriori algorithm, classical algorithm for the discovering of one-dimensional Boolean association rules, is chosen to apply. Moreover, Apriori algorithm is ameliorated in order'to better serve the purpose of Data Mining in the Traditional-Chinese Medicine prescription database: MApriori algorithm -to discover multi-dimensional Boolean association rules, and WApriori algorithm - to discover weighted Boolean association rules that assign a weight to each item of the database according to its importance to the generation of rules and take the weights into consideration when generating the association rules - are brought forward. To discover clusters, agglomerating hierarchical algorithm is used. To enhance the quality of clustering, RatioD Distance is used in the place of classical Euclid Distance.This paper is based on the development of software "Traditional-Chinese Medicine Prescription Analysis System" which was developed by the cooperation of the Laboratory of Neural Networks and Information Technology of Southwest Jiaotong University and TheLibrary of Chengdu Chinese-Medicine University. This software is a data-mining tool for Chinese-Medicine database whose function includes: 1) to help the Chinese-Medicine researchers discover hidden principles and trends in the Chinese-Medicine prescriptions; 2) to help the Chinese-Medicine doctors make more scientific and standard decisions when giving prescriptions.