The Studying Of Mining Association Rules In Database And The Application In Stock Market Block's Moving For Each Other

Mining previously unknown and potentially useful information and knowledge from large databases is an urgent need posed by the explosive growth in data. Consequently, data mining technology has emerged for this purpose. Association rule is an important kind of knowledge patterns that are discovered by data mining.Now the research on data mining becomes a hot topic. Because of its clear description of the relations with certain level among the realistic objects, which is very practical, association rule attracts a large quantity of researchers in many different fields. In this paper, we put emphasis on the algorithms to mine association rules and their implementations with reference to the researching condition of association rule and the latest tendency in these topics. I first introduced the basic concept , research methods and common methods of data mining technology; Secondly on the basis of the research condition of association rules presently and the common process of mining association rules , the paper does a further studying and discusses main studying direction; Thirdly , it analyses several kinds of basic algorithms in mining association rules and points to the shortcoming in them. It puts forward two algorithms which are data mining algorithm based on prefer linking list andfrequent pattern growth tree to produce association rules so as to improve the mining efficiency which is reduced by so many indexing database. It proves their effectiveness of these two strategies in theory. The two algorithms only scan database two times. They need not produce excessive candidate large item sets, so that they not only improve the running efficiency but also save the space of interior storage. During producing the rules, we take effective methods to increase it' s speed and as large as probably to reduce the times of divide. The whole algorithm has been improved in the two sides so as to largely increase the algorithm' s speed. In this thesis we also put the theory into the reality of the stock blocks. After running, we receive a pleasing result. At last I draw a conclusion from the whole paper and look forward to the future' s blueprint in this field.