Re: what is data mining?

Anthony Cowden (cowden@sonalysts.com)
Tue, 30 Mar 1999 22:01:51 +0200 (MET DST)

A VERY good definition! I like it!

We would like to introduce folks to the use of fuzzy logic in data ANALYSIS.
It is our contention that the spectrum running from data retrieval (i.e.,
SQL-based queries) to data mining (see definition below) leaves room for data
ANALYSIS. The use of fuzzy logic allows you to determine the degreee to which
a record meets the semantic intent of your query (with SQL, it either meets it
or it doesn't; very binary!).

See the URLs below for more info...

Jerry Clough wrote:

> I'm a bit of a purist as to defining data mining, and prefer definitions
> along the line:
>
> "The automated discovery of previously unsuspected
> relationships between items of data held in a database"
>
> IMHO this excludes the following types of activity,
> which are often described as data mining:
>
> * Ad hoc database queries (e.g., SQL against an RDBMS)
> by a knowledgable business analyst. Its not automated,
> and usually the analyst has a pretty good hunch that a
> relationship exists.
>
> * OLAP type queries against a multi-dimensional schema
> (whether in an OLAP tool or against a star schema). Usually
> someone has imposed a highly structured (and therefore
> pretty well understood) view of the relationships between data
> items.
>
> * Probably some others I have a gripe about!
>
> The classic data mining example is the "beer & diapers/nappies"
> story. This goes along the line: using a data warehouse a West
> Coast retailer noticed that diaper sales were sharply higher late
> on Friday afternoons. Further examination showed that the
> purchasers were men. Positioning high-margin beer brands
> opposite the diapers got much higher sales from these customers.
> The logic was that wives were phoning their husbands and saying
> "I'm worried we may run out of diapers at the weekend can you
> pick some up from the store". As these were husbands with very
> young children they were probably missing Firday night beer drinking
> and would therefore make the impulse purchase.
>
> You need automation to pick this kind of thing up because a retailer
> might have 20,000 lines in store at any one time, and the potential
> cross-selling matrix is vast.
>
> Suggest you look at the following sites for a much better o'view
> than I could hope to give you:
>
> http://www.cs.bham.ac.uk/~anp/TheDataMine.html
> http://www.kdnuggets.com/
>
> Jerry Clough

--
**********"Leadership is about being out front..." ***********

Anthony Cowden, Manager, Fuzzy Systems Solutions

Sonalysts, Inc.

Fuzzy Systems Solutions: http://www.sonalysts.com/fuzzy.html

Fuzzy Query (TM): http://www.sonalysts.com/fq.html

############################################################################ This message was posted through the fuzzy mailing list. (1) To subscribe to this mailing list, send a message body of "SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at (2) To unsubscribe from this mailing list, send a message body of "UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com" to listproc@dbai.tuwien.ac.at (3) To reach the human who maintains the list, send mail to fuzzy-owner@dbai.tuwien.ac.at (4) WWW access and other information on Fuzzy Sets and Logic see http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info (5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html