RE: DM: Datamining tools ... black boxes...

Clay said:

	> The definition quoted by Warren refers to an electric
	> circuit with unknown means of operation, but *someone*
	> must know how it works or how could they have built it?
	> When a person calls something a black box, it's because
	> *that person* doesn't know how it works, not because
	> *nobody* knows how it works.

I'm much more comfortable with Warren's statement, and I'd be
unhappy if the notion that "black box" means "the user doesn't
understand it" in the "new and emerging" jargon of DM.

Warren's comment meant that the BEHAVIOUR and
PERFORMANCE of the "box" were accessible, but that the
OPERATIONAL DETAILS were essentially private - whether you
could make sense of the internals or not was irrelevant as they
were out of sight.

If a user wants to know the answers and how reliable they
are, "black box" is fine. If the user wants to know how the
answers were obtained in detail, that's "white box".

However, if the user wants to know how to interpret or explain
the results, that's nothing much to do with "boxes". Different
tools give different scope for interpretation - which should be
tempered by GOOD business/domain analysis, and "deep
thought". As far as I know, DM does not provide "deep thought",
but it can provide things to "think deeply" about.

Cheers, Tom - who just realised why his neural models have
converged slower with better data...

Dr Tom Osborn
Director of Modelling
NTF
Decision Support Consultants
Level 7, 1 York Street
SYDNEY NSW 2000
AUSTRALIA
phone:	+61 2 9252 0600
fax:	+61 2 9251 9894

• Prev by Date: DM: Job Opportunties at Durham
• Next by Date: DM: Position at Data Mining Group, Siemens Corporate Research
• Prev by thread: RE: DM: Datamining tools ... black boxes...
• Next by thread: RE: DM: Datamining tools ... black boxes...
• Index(es):
  • Date
  • Thread