Title: Statistical techniques for rough set data analysis
Authors: Günther Gediga , Institut für Evaluation und Marktanalysen; Brinkstr. 19; D-49143 Jeggen; Germany
Ivo Düntsch , Dept of Computer Science , Brock University , St Catherines, Ontario, L2S 3A1, Canada
(Equal authorship implied)
Status: In: Rough Set Methods and Applications (Ed: Lech Polkowski, Shusaku Tsumoto, Tsau Young Lin), Physica Verlag (2000), 545-565
Abstract: Although designed as a structural (in particular, a non statistical) approach to data analysis, application of rough set data analysis only makes sense, if some basic statistical assumptions are observed. We describe how the application of these assumptions leads quite naturally to
  • Statistical testing schemes for the significance of inference rules,
  • Entropy measures for model selection, and
  • A probabilistic version of RSDA.

View technical report version