The Association Rules Mining Algorithm Based On Genetic Algorithm And An Improved Interest Model

The Internet has changed people's ways of thinking and behavior, in this information society the human has accumulated vast amounts of data, and how to use these data to dig out the "gold" knowledge has become more and more important.Association rules is an important research content in the field of Data mining,and the reason for researching the association rules is to find out the correlation between items and items from a large amount of data. This paper first introduces the basic concepts and properties about association rules, and then introduces the apriori algorithm in detail, a classical algorithm of association rules, then introduces the genetic algorithm briefly, including the basic terms and operation process of genetic algorithm, which is one of the important methods of data mining.This paper analyzes that the mining method which based on the traditional framework which includes support and confidence may dig out some useless and even misleading rules and some useful association rules may can not be digged out.So the people introduce the interest to improve the traditional framework because it can tell us the rules' degree of interesting and eliminate some useless or misleading rules.Firstly, this paper proposes an improved interest model, and then proves the interest model is better than the common interest model in theory and experiment.This paper analyzes the natural advantages of genetic algorithm in association rule mining., summarizes the current situation and problems of using genetic algorithm on association rules mining, and discusses the idea of combining genetic algorithm with interest for mining association rules. Finally, a new association rule mining algorithm based on the genetic algorithm and the improved interest model is proposed,which named an association rules mining algorithm which based on genetic algorithm and improved interste model.The paper demonstrates the proposed algorithm can more out uninteresting rules throw compare with other two existing association rules mining algorithm,so the algorithm proposed in this paper is efficient and has certain application value, which can be applied to the research and practice.