An intelligent framework (O-SS-E) for data mining, knowledge discovery and business intelligence

KDD at its inception offered much to support the building of decision support systems, and ´business intelligence´, through its data mining and data warehousing technologies. However, the reality seems to not to have matched the hopes. The DM models used in DW are largely trivial, and the data-structures used in DM are trivial: a schism seems to have developed between DM and DW. Furthermore there seems to be no coherent DM methodology to guide the choice of models and their appraisal. It is proposed that linked ontologies of data, and models provide such guidance, and a theory of knowledge (an epistemology) needs to be developed in order to enable appraisal of discovered knowledge. Between these front-end and back-end suggestions for the development of a coherent KDD methodology it is also suggested that DM toolkit needs major overhaul by the inclusion of the extensive use of search and sampling techniques at the strategic DM level.