The importance of a "Data Mining oriented analysis phase" in a data warehousing project methodology

Classical two or three level architectures proposed so far for data warehouses show some drawbacks when adopted to work over large numbers of heterogeneous operational sources. In the application context under consideration, having a suitable architecture may not be enough for design purposes. Indeed, data warehouse and data mining architectural substrate design in very large operational environments can be a quite hard problem to be attacked with traditional manual methodologies. Therefore, the purpose of this paper is to provide an alternative four-level data warehouse architecture (DMOIS) and a methodology phase, partially data mining based and oriented.