Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!EU.net!sun4nl!news.nic.surfnet.nl!spider.research.ptt.nl!sun005!klooster
From: klooster@sun005.research.ptt.nl (Sytse Kloosterman)
Subject: [Q] Knowledge Discovery in DBs tasks
Message-ID: <1994Nov2.100233.29957@news.research.ptt.nl>
Sender: klooster@sun005 (Sytse Kloosterman)
Nntp-Posting-Host: sun005.research.ptt.nl
Organization: PTT Research, Groningen, The Netherlands
Date: Wed, 2 Nov 1994 10:02:33 GMT
Lines: 32

Hi,

I'm working on Knowledge Discovery in Databases (KDD) which is also known
as data mining. Based on known literature, we distinguish four types of
KDD tasks:
* clustering,
* class description (which can be divided in summary and discrimination),
* dependency analysis,
* deviation detection.

When one tries to interrelate these tasks or when one tries to formulate
their differences, it turns out that this is not easy to do. A specific
task of one catergory can sometimes also be seen as belonging to
another category.

Well, here are my questions: 
* what are the precise definitions of the above mentioned tasks,
* how are they related, and
* what are the differences?

To give a first attempt, I think that clustering is the process of grouping
objects. Objects are grouped because they are similar enough. Now, based
on these clusters, the other tasks can be viewed as functions defined on
one or more clusters returning a pattern which is a representation of a
summary of the cluster, a trend in a series of clusters, etc.

Please shoot!


Thanks & regards,

Sytse.
