Research of Data Mining Service Application and Implementation of Analysis Service

In this thesis, we studied the theory of OLAM which combined data mining with On-Line Analytical Processing (OLAP). This thesis presents a kind of distributed data mining service architecture aiming to resolve the problems that appear widely in the current data mining tools, such as low level of resource sharing, functional coupling too closely, difficult to reuse software and so on. On the basis of service architecture and its Hierarchical thinking, the overall structure was designed and analyzed specifically; all levels of design were also illustrated in details. Finally, OLAP service on sales data was implemented based on Microsoft SQL Server 2005 and sales forecast project was built based on cluster analysis algorithm as well. It has made a foundation for customer data analysis and sales rules mining, and provided a scientific decision support method for the analyst in this area.