Seminář: Association Rules: Postprocessing of Hypotheses

In mining association rules from databases is far from easy the challenge to go through the phase of postprocessing the discovered association rules. Several techniques for advanced postprocessing were described [1], [2]. The presentation contains additional characteristics of the set of the discovered association rules, that should be explored by user. New method for association rules postprocessing will be presented. The method is based on transformation of the association rules into text documents and on two independent clustering processes. The aim of the new method is to get credible and user easy to understand description of set of the discovered association rules. [1] G. Dong, J. Li. Interestingness of Discovered Association Rules in terms of Neighborhood-Based Unexpectedness. Research and Development in Knowledge Discovery and Data Mining, Proc. 2nd Pacific-Asia Conf. Knowledge Discovery and Data Mining. Springer 1998. [2] H. Toivonen, M. Klemettinen, P. Roikainen, K. Hatonen, H. Mannila. Pruning and Grouping Discovered Association Rules. In Workshop Notes of the ECML-95 Workshop on Statistics, Machine Learning, and Knowledge Discovery in Databases, pp. 47 — 52, Heraklion, Greece, April 1995. MLnet.