CALL FOR PAPERS
 Reliability Issues in Knowledge Discovery

A Special issue of the International Journal

Data Mining and Knowledge Discovery

Guest Editors

Dr Honghua Dai, Deakin University, Australia

Prof Jiawei Han, Univ. of Illinois at Urbana-Champaign, USA

[Description]

With the rapid development of the data mining and knowledge discovery, a key issue which could significantly affect the real world applications of data mining is the reliability issues of knowledge discovery. It is natural that people will ask if the discovered knowledge is reliable. Why do we trust the discovered knowledge? How much can we trust the discovered knowledge? When it could goes wrong. All these questions are very essential to data mining.

One of the essential requirements of data mining is validity. This means both the discovery process itself and the discovered knowledge should be valid. Reliability is a necessary but not sufficient condition for validity. Reliability could be viewed as stability, equivalence and consistency in some ways.

This special issue of the international journal on data mining and knowledge discovery on the reliability issues of Data Mining and Knowledge Discovery will focus on the theory and techniques that can ensure the discovered knowledge is reliable and to identify under which conditions the discovered knowledge is reliable or in which cases the discovery process is robust. In the last 20 years, many data mining algorithms have been developed for the discovery of knowledge from given data bases. However in some cases, the discovery process is not robust or the discovered knowledge is not reliable or even incorrect in certain cases. We could also find that in some cases, the discovered knowledge may not necessary be the real reflection of the data. Why does this happen? What are major factors which affect the discovery process? How can we make sure that the discovered knowledge is reliable? What are the conditions under which a reliable discovery can be assured? These are some interesting questions to be investigated.

[Scope and Topics]:

The topics of the special issue will include but not be restricted to the following:

• The theories on reliable knowledge discovery
• Reliability measurement criteria of knowledge discovery
• General reliability issues on knowledge discovery
• Domain specific issues on knowledge discovery
• The criteria that can be used to assess the reliability of discovered knowledge.
• The conditions under which we can confidently say that the discovered knowledge is reliable.
• The techniques which can improve reliability of knowledge discovery
• Practical approaches that can be used to solve reliability problems of data mining systems.
• The theoretical work on data mining reliability
• The practical approaches which can be used to assess if the discovered knowledge is reliable.
• The analysis of the factors that affect data mining reliability
• How reliability can be assessed
• In which condition, the reliability of the discovered knowledge is assured.

[Submission and Review]

All submissions will be reviewed on the basis of interestingness, relevance, originality, significance, usefulness, soundness and clarity by at least two experts independently.  Submit papers via the journal website, selecting the special issue under Article Type.

Important Dates:

Extended deadline for paper submission: 30^th, April 2007
Author notification: 20^th, June 2007
Camera-ready copy due:  15^th, July 2007