Re: No brainer of the day

In article <3ughsf$26t@cssun.mathcs.emory.edu>, Mark.Denham@bbc.co.uk (Mark Denham) writes: |> |> Ok all, |> |> I'm tired of hearing the terms data mining and data warehousing without |> fully understanding what they mean. I also get the impression that their are |> different views of what they mean being banded about. Data mining means finding patterns in large volumes of data. Ideally one would have an AI inference engine of some sort to do the work. In practical terms it means a whole lot of queries w/ no pre-determined likelyhood of finding anything useful. So mining is an activity. Data warehousing is a suite of processes for administering a large volume of time-phased data. Within this suite one might find a data warehouse, an operational data store, middleware, a metadata repository, and a range of client query tools. The ODS is most often used to produce list style reports and possibly some level of OLTP activity. The warehouse serves a majority of people doing routine, periodic queries. Inmon calls them farmers - they till the same ground. And a few explorers - or miners as you will, who range freely. The warehouse may be instantiated in several forms: traditional RDBMS and/or Multi_dimensional engines, with the later form better for data mining activities. The jargon is complex; the ideas are *rapidly* evolving. Hope this helps. Dave Nelson ___________________________________________________________ Hewlett Packard email: daven@corp.hp.com ICBD fax: (415) 852-8312 1501 Page Mill Rd. phone: (415) 857-2902 Palo Alto CA. 94304 ___________________________________________________________