Knowledge discovery in large information resources

Aim of project
Knowledge discovery in large information resources is one of the most extensively developed topics in the fi eld of contemporary computer science both from the theoretical and practical stand point. Although unquestionable achievements have been made in this area, still new problems arise mainly related to the effi ciency of data mining algorithms and the discovery of essential pieces of knowledge in the abundance of patterns. The problems are particularly apparent in text and spatial databases. Despite recent achievements in the methods of discovering and using condensed knowledge representations, works on eff ective representing and discovering generalized knowledge are not available in literature.

In particular, mining essential knowledge in the form of formulae with negation remains an open problem.
The aim of the project is:
The development of new methods of discovering knowledge as formulae with negation, which allow construction of a representation from which all essential dependencies and relations, held in a database, can be derived;
Experimental verification of these methods, including comparison with known methods applied to discover knowledge in large databases;
Adaptation of new methods for mining text and spatial databases.

Expected results
The expected results of the project shall include: software prototype, journal publications, conference presentations, Ph.D. theses. The project results shall be useful for theoretical and practical works related to data/knowledge mining systems and, in particular, for creating integrated and functionally homogeneous environments for analyzing large information resources.