Research On Multi-dimensional Association Rules Mining

With the rapid development of database technology and its wide application, the amount of data is accumulated in the database. There are many important messages behind the increased data. People hope to do the higher level analysis on the scientific research, business decision or Enterprise management based on the vast amount of data, so as to obtain the valuable infomation and provide the gist for the management and decision-making.Data Mining is to pick-up implied massages that people do not know in advance from the incomplete, noise, fuzzy data, but is potentially useful information and knowledge process.The main objective of this paper is to study an effective way to found the association of multidimensional data. Multi-dimensional data analysis which based on the data warehouse and OLAP is to analysis multi-dimensional data by the method of statistical analysis, and multi-dimensional association mining algorithm is to analysis multidimensional data by the method of knowledge discovery. This paper was studied from these two aspects.In this paper, firstly, the author does the study of multi-dimensional data analysis which based on data warehouse and OLAP. The main part is data warehouse modeling, and the method of using OLAP analyzing multidimensional data in the data warehouse. On the basis of the study, the author makes a model of"Analysis System which based on warehouse and OLAP to analysis the employment situation of college graduates data ", and adopt slice,dice, roll-up and drill-down to analyses the employment situation of college graduates;Secondly, in this paper, the author makes a study about association mining method and the development situation. On the basis of it, the author provides the multi-dimensional association rule mining algorithm based on the SQL. It uses CUBE operators and combines with SQL powerful queries function, and it can mine a certain number of data sets efficiently. In order to make effective mining about greater data sets, algorithm is improved .It divides the large data set into smaller ones. In this paper, the algorithm and design idea are discussed in detail, and gives some of the algorithm test results. Analysis and experiments show that algorithm is simple and easy to realize, without scanning mining data sets many times, and have high mining efficiency.Finally, in order to make the users easily use the multi-dimensional association mining algorithm to mine multi-mention data, the author designs multidimensional association mining tools based on the powerful data processing of SQL Server database and C # object-oriented functional integration .