Study On Association Rules Based On Fuzzy Set Theory

With the revolution of IT, huge datasets are deposited. Traditional methods or algorithm are difficult to deal with these datasets. Data mining methods are applied in the objective. Association rule algorithms is very important research aspect in the data mining field. Association rule is a kind of knowledge patterns which describes a rule that different attribute of one transaction appeared in the same time. This paper describes the application architecture of data mining, the mining theory of association rules and the mining algorithms of association rules. In the application architecture of data mining, this paper analyzes the basic processing phases of data mining or KDD, and describes the components and functions of a data mining application system. We introduce the existing algorithms of association rules-----Apriori, analyze the disadvantage of this algorithm. We use the Fuzzy theory to discover associate rules from the datasets consisting of non-boolean attributes.. With fuzzy theory, it is possible to summarize and abstract the mining dataset for association rule discovery which leads association rules to be discovered and also extend the application of association rules. we proposed a fuzzy association rule expressed by the fuzzy concept, and extended the manifestation and application of the definite association. At the same time ,we discussed the special character of fuzzy association rules; and then based on this , we applied a solution ----MFAR(Mining Fuzzy Association Rule).Moreover, we explain the interesting association rules by means of decision tree generated by See5.