Fuzzy Association Rule And Its Application

Data mining is a new subject which arises with the development of artificial intelligence and database. Association rule is one of the important research tasks, and also one of the key technologies in data mining. Its purpose is to find out the association between all items in business database. Association rule plays a very important role in decision system, expert system and aptitude information system in every aspect. Many experts have been making research on associaton rule, and have made big progress based on original methods. They also find out several new methods with respect to the knowledge of many other fields.This paper gives out an analysis of the definitions of association rule in the existing books and papers. Then it points out the cause of the difference, normalizes the description and makes out new descriptions from the angle of set and possibility, which lays a foundation on the further research of the theory and application of the association rule.Then we use the grey association rule to make empirical analysis. This theory is often used to analyze the feasibility of a project, but in this paper it is used to find out the relationship between the questions in a questionnaire by analyzing the results. After that, a membership fuction is given out by considering the feature of the original data, which is used to analyze the result given by the grey association rule. In the end, we use the two rules together to analyze the industrial structure of Xiamen City. The grey association coefficient is used as part of the weight in fuzzy association rule to find out fuzzy weighted association rule. Finally we gives out some suggestions according to the anyalysis result.The main contributions of this dissertation are that firstly, we points out the inconsistency and the possible confusion that it leads to based on the summerization of the existing expression of association rule, and normalize the description ; secondly,gives out a membership function considering the feature of data; thirdly, we link the two methods together, the grey association coefficient is used to be part of the weight in fuzzy weithted association rule.