The Design and Implementation of Chinese Medicine Groups Mining System based on Genetic Algorithm
Abstract
This paper researches the compatibility of chinese medicine prescriptions by data mining techniques. The main contributions include: analyzes the problems in the association rules, and introduces the concept of the bidirectional association rule; presents the foundation principle of genetic algorithm(GA), and studys the application of GA in the data mining; converts chinese medicine prescriptions database to a bitmap matrix, which greatly enhances the efficiency of search; develops a chinese medicine groups mining system based on GA. The paper is organized as follows: Section 1 introduces the background and significance; Section 2 sets forth the basis of the relevant theories; Section 3 proposes the design project of the system; Section 4 detailedly shows the implementation of the system, including the implementation of bitmap matrix, the individual coding method, the design of fitness function, rules of the extraction, genetic operations. Section 5 gives a test of the system on the prescriptions database about spleen and stomach, and analyzes the results. It is proved that this system can find important and significant Chinese Medicine Groups from the prescriptions database, and is meaningful for the research of Chinese medicine.
Key words: Data mining; Confidence; Bidirectional association rule; Genetic algorithm