Affiliations: Department of Expert Systems and Artificial
Intelligence University of Information Technology and Management 35-225
Rzeszow, Poland. E-mail: {pblajdo,zhippe,tmroczek,mknap}@wsiz.rzeszow.pl | Department of Electrical Engineering and Computer
Science, University of Kansas Lawrence, KS 66045, USA and Polish Academy of
Sciences 01-237 Warsaw, Poland. E-mail: jerzy@ku.edu | Department of Distributed Systems, University of
Information Technology and Management 35-225 Rzeszow, Poland.
E-mail: lpiatek@wsiz.rzeszow.pl
Abstract: We present results of extensive experiments performed on nine data
sets with numerical attributes using six promising discretization methods. For
every method and every data set 30 experiments of ten-fold cross validation
were conducted and then means and sample standard deviations were computed. Our
results show that for a specific data set it is essential to choose an
appropriate discretization method since performance of discretization methods
differ significantly. However, in general, among all of these discretization
methods there is no statistically significant worst or best method. Thus, in
practice, for a given data set the best discretization method should be
selected individually.
Keywords: Rough sets, discretization, cluster analysis, merging intervals, en-fold cross validation, test on the difference between means, F-test, sign test
