Nautilus Systems, Inc. logo and menu bar Site Index Home
News Books
Button Bar Menu- Choices also at bottom of page About Nautilus Services Partners Case Studies Contact Us
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [Subscribe]


From: Ivan Bruha
Date: Tue, 23 Mar 1999 14:06:12 -0500 (EST)

                  C A L L   F O R  P A P E R S

From Machine Learning to Knowledge Discovery in Databases 

                 :::::A WORKSHOP WITHIN:::::

                      (( ICML-99 ))

                Bled, Slovenia, 27-30 June 1999

This workshop addresses an important aspect related to Knowledge Discovery in Databases (KDD) or Data Mining (DM) and Machine Learning (ML) in pre-processing and analyzing real-world data.

Knowledge Discovery in Databases has become a very attractive discipline both for research and industry within last few years. Its goal is to extract "pieces" of knowledge or "patterns" from usually very large databases. One of its components is an inductive process which induces the above "pieces" of knowledge; usually it is machine learning. However, most of the machine learning algorithms require more or less prepared data in a reasonable format. Therefore, some preprocessing routines as well as postprocessing ones should fill up the entire chain of data processing.

The data which are to be processed by an algorithm are usually noisy and often inconsistent. Many steps are involved before the actual data analysis starts. Moreover, many ML systems do not easily allow processing of numerical attributes as well as numerical (continuous) classes. Therefore, certain procedures have to precede the actual data analysis process.

Second, a result of an ML algorithm, such as a decision tree, a set of decision rules, or weights and topology of a neural net, need not be perfect from the view of custom or commercial applications. It is quite known that a concept description as a result of an inductive process has to be usually post-processed. Post-processing procedures usually include various pruning routines, rule quality processing, rule filtering, rule combination, or even knowledge integration. All these procedures provide a kind of "sym- bolic filter" for noisy, imprecise, or "non-user-friendly" knowledge derived by an inductive algorithm.

Thus, the pre- and post-processing tools always help the DM algorithms to investigate databases as well as to refine the acquired knowledge. Usually, these tools exploit techniques that are not genuinely logical, e.g., statis- tics, neural nets, and others.

These reasons let us to launch this workshop. We would be pleased to accept papers from the following areas:

o Mapping data
o Scaling learning algorithms to large datasets
o Handling noise
o Processing of unknown attribute values
o Discretization/fuzzification of numerical attributes
o Grouping of values of symbolic attributes
o Consistency checking
o Attribute (feature) selection and ordering
o Constructing new attributes
o Transforming attributes
o Processing of continuous classes
o Interpretation and explanation
o Evaluation
o Knowledge combination and integration

This workshop provides an opportunity for researchers to learn about the challenges and real problems in development and applications of machine learning techniques.


Ivan Bruha
McMaster University               
Dept. Computing & Software                 Phone: +1-905-5259140 ext 
Hamilton, Ont.                             Fax:   +1-905-5240340
Canada  L8S 4K1                            Email:

Marko Bohanec
Institut Jozef Stephan
Jamova 37
Ljubljana, Slovenia                        Email:

Program Committee:

A. (Fazel) Famili
Editor-in-Chief, Intelligent Data Analysis
Institute for Information Technology       Phone: +1-613-9938554
National Research Council of Canada,       Email:
Montreal Rd, Ottawa, Canada  K1A 0R6

Ivan Bruha, McMaster University, Hamilton, Canada

Marko Bohanec, Institut Jozef Stephan, Ljubljana, Slovenia

Stan Matwin, University of Ottawa, Information Technology and 
Ottawa, Ont., Canada  K1N 6N5
Email: stan@site.uottawa.ca5

Gholamreza Nakhaeizadeh, Daimler Benz AG, Germany

Igor Kononenko, Ljubljana Univ., Ljubljana, Slovenia

Petr Berka, Laboratory of Intelligent Systems, University of 
Prague, Czech Republic

W.F.S. (Skip) Poehlman, McMaster University, Hamilton, Canada

Organization Notes:

There will be one invited talk on the workshop which will survey the given topic as well as introduce own research.

About upto 10 accepted papers will be presented (each 15-20 min). If there is a larger interest, then some papers might be accepted as posters. Maximum size is 10 (ten) pages.

Attendance is not limited to paper authors. However, in order to get an early estimate of the possible attendance, we would appreciate an informal note about your intention to attend.

A panel session at the end of the workshop will summarise what has been learned from the workshop and will identify future directions.

>>>>>> Please note that authors of the best papers will be invited >to submit an extended version of their papers to the Intelligent Data Analysis Journal (, or even a special issue of the journal regarding this topic might be published.


Submit your paper either by regular mail or by Email to I. Bruha (see the address above). If you use Email, then the Postscript format would be the most suitable one.

Important Dates:
Deadline for submission: 30-Mar-99
Notification: 30-Apr-99
Camera-ready copy: 15-May-99

[ Home | About Nautilus | Case Studies | Partners | Contact Nautilus ]
[ Subscribe to Lists | Recommended Books ]

logo Copyright © 1999 Nautilus Systems, Inc. All Rights Reserved.
Mail converted by MHonArc 2.2.0