Research And Improvement Of Apriori Algorithm In Association Rules

With the development of science and technology, people can collect data with more convenient way.The rapid growth of database makes people urgently need a technique to deal with these data.Data mining is produced under such background. The data mining analyzes, extracts or mines the knowledge from data which is previously unknown, novel, and potentially valuable for decision-making.Data mining is the most important step in the KDD process.Data mining common function has many kinds,such as clustering, classification and association rules, etc..This paper mainly introduces the association rules among the.paper mainly introduce some basic concepts of the association rules and some of the basic methods of mining association rules.In this paper,an analysis and study is presented on association rule mining od data mining. And then improved the classic Apriori algorithm bases on the boolean matrix. The main researches of this thesis are as follows:(1) A general research on the basic theory and classification of data mining is done, and the basic idea and classical algorithms of association rule mining are analyzed and researched primarily.(2) Several existing efficient algorithm are introduced in this paper, and then analyzed the performance of this approved algorithm. After that, the advantages and disadvantages are compared.(3) On the basis of analyzing association rule, a algorithm for association rule is put forward based on matrix. The new algorithm maps the database into a0-1matrix, computing directly on the matrix and avoiding scanning the database repeatedly. And also extended and utilization the properties of the Apriori, the matrix will be compressed more thoroughly. At the end of the paper, a theoretical analysis and experiment will been givenThe new algorithm introduced the idea of vector inner product, the performance ascendant of the algorithm is proven by theory analysis and experiment.