The Study Of Multidimensional Association Rule Mining Method Based On Data Cube And Drill Technology

Along with DBMS's wide application in various domains, The DB application technique has rapid progress. The data are more and more. The association rule mining is a very useful model in the techniques of data mining. Many researchers are processed a large numbers of studying for the problems of association rules mining. Their jobs are that the intrinsic algorithm was optimized, such as random sample, and collateral thought, etc. so that we can improve the efficiency of algorithm which is used mining rules, and generalize the applications of association rules. But single Boolean classical algorithm Apriori's core is the successive methods based the theory of frequent items. Apriori's character determinates it. We need search after the new methods of mining association rules, so as to avoid several bugs of frequent items. Along with the development of databases and data warehouses, it is very important that how to use the data carries through the data mining. OLAP's technique based on the data analytic necessaries constructs all kinds of data cube in original data, and handles the data cubes, then returns the results to users. We know that there is a very consanguineous relation between data warehouse and data mining. Data mining is a sort of support of decision-making technique, what based on data warehouse. Data mining is a process of filtering potential useful information in large numbers of data. The process disports 4 steps in data warehouse system, data choice, data transform, data mining, and analyzing result. It is the reason because data warehouse adsorbs the strongpoint of OLAP's drilling; we induct using dimensional association rules mining in data warehouse. At the same time, we spurn the burden of scanning database time after time and clipping frequent items, which stems from Apriori's characters. To avoid adds the computer's burdens at adjusting dimensional levels, soothe algorithm's efficiency is advanced. At present, the studying of multidimensional association rules mining method based on data cube focus on the amelioration different of algorithm of data cube of different densities. There are many achievements for low-density data cube. I find it can realize in my gathered information, that mining multidimensional data's association rules use drilling. I expect that the mining is effective, which don't use Apriori's characters. The first chapter introduces the basic conceptions. There are association rules and it's species, multidimensional association rules and its types, data cube and it's structure, OLAP technique, drill-up, drill-down, and multidimensional data schema of data warehouse. The second chapter introduces multidimensional association rules mining based on data cube, and Apriori_cube algorithm. It comes from the other's conclusions. Firstly, it describes the process the multidimensional association rules mining based on data cube. Secondly, it separates explaining each step, and pictures the Apriori algorithm. Thirdly, it analyzes algorithm from time and efficiency. The third chapter explains the amelioration of Apriori_cube. It inducts the algorithm's core, and defines the algorithm of Apriori_cube_dimenlevel, then analyzes the new algorithm. The fourth chapter is experimentation. It used SQL Server 7.0, discusses the new algorithm's intention. The fifth chapter is conclusion. It summarizes the achievements and insufficient. We view the whole article; you know that mining multidimensional association rules need not only the theories, but also the practices, especially, OLAP's applications. This algorithm enhances the contra pose of adjusting. It is more accord with the users require. The new algorithm is meaning and impression from analyzing experimentations.