6th International Conference on
Knowledge-Based Intelligent Information Engineering Systems & Allied
Technologies
16, 17 & 18 September 2002, Podere d'Ombriano, Crema, Italy


Special Session
Classification in Knowledge-Based Systems


Topic :

This session aims at studying methods of classification in the
framework of knowledge-based systems.  Traditionally, classes involved
typically have a hierarchical organization where the subclasses have
discriminating features from the superclasses.  Classification is then
a top-down process during which the system tries to match data to
classify with classes at each level of the hierarchy until th= e
matching is not satisfying any more.  Another way of viewing
classification is Clancey's one who has defined heuristic
classification as a problem solving method which consists in
establishing a relation between data that are associated with a
specific problem and solutions already defined and present in a
knowledge base.  Heuristic classification is composed of three steps :

- a data abstraction step, which transforms the data that describe
the problem into data abstractions, by possibly using rules, 
- a heuristic matching step during which the previous data are associated
with a generic hypothesis, 
- a step which allows to specialize the solution hypothesis and where the 
data concerning the problem are used to refine this hypothesis.

Some "classical" classification systems adopt an hybrid approach by using
heuristic components.

Contributions to this session are expected to deal with those two views of
the classification process :
- - for the classical approach to classification, sub-topics of interest woul=
d
be : new methods of matching using prior knowledge, new strategies for
exploring hierarchies in order to classify data, particular types of
knowledge representation which allow to refine the matching step, etc=8A
- - for the heuristic approach to classification, sub-topics of interest woul=
d
be : new methods for generalizing specific data, definition of innovating
heuristic matching methods, improvement of the refining step, etc=8A

The previous list is not exhaustive, provided that the contributions are
relevant to the topic description.

Both theoretical papers and papers describing practical experiences will be
welcome.


Instructions for Authors :

Only electronic copies of the papers in Microsoft Word, PDF or Postscript
forms are acceptable for review purposes and must me sent to the session
chair.  However, please note that you will be required to send hard copy of
the final version of your paper, if it is accepted; electronic submission o=
f
final papers is not allowed.

Papers must correspond to the requirements detailed in IOS Instructions for
the Preparation of a Camera-Ready Manuscript. The maximum length of papers
is five pages, prepared according to the IOS instructions.  Longer papers
will be subject to an Additional Page charge.

All papers must be presented by one of the authors, who must pay fees.


Publication :
The session papers will be published in the Conference Proceedings (IOS
Press,  Amsterdam, the Netherlands).

Extended versions of selected papers will be considered for publication in
the KES Journal (International Journal of Knowledge-Based Intelligent
Engineering Systems).


Important Dates :
Deadline for submission intention : December 15, 2001
Deadline for submission (Receipt of papers) : February 10, 2002
Notification of acceptance : April 10, 2002
Camera-ready papers by : May 10, 2002


Session Chair :
Dr. Colette Faucher
Associate Professor of Computer Science
Universit=E9 d'Aix-Marseille III
Facult=E9 des Sciences de Saint-J=E9r=F4me
DIAM-IUSPIM
Avenue Escadrille Normandie-Niemen
13397 Marseille Cedex 20
FRANCE
Phone : (+33) 4 91 05 60 58
Fax : (+33) 4 91 05 60 33
E-mail : [EMAIL PROTECTED] or [EMAIL PROTECTED]
Website of the session :
http://perso.wanadoo.fr/colette.faucher/kes2002-classification.html

Reply via email to