OMARS: The Framework of an Online Multi-Dimensional Association Rules Mining System

Recently, the integration of data warehouses and data mining has been recognized as the primary platform for facilitating knowledge discovery. Effective data mining from data warehouses, however, needs exploratory data analysis. The users often need to investigate the warehousing data from various perspectives and analyze them at different levels of abstraction. To this end, comprehensive information processing and data analysis have to be systematically constructed surrounding data warehouses, and an on-line mining environment should be provided. In this paper, we propose a system framework to facilitate on-line association rules mining, called OMARS, which is based on the idea of integrating OLAP service and our proposed OLAM cubes and auxiliary cubes. According to the concept of OLAM cubes, we define the OLAM lattice framework that exploit arbitrary hierarchies of dimensions to model all possible OLAM data cubes.