Two rough set based feature ranking methods
- Prelegent(ci)
- Sebastian Stawicki
- Termin
- 7 października 2011 15:45
- Pokój
- p. 5820
- Seminarium
- Research Seminar of the Logic Group: Approximate reasoning in data mining
I briefly overview two rough set based feature ranking methods along with a new algorithm called Rough Attribute Ranker (RAR). RAR assesses the usefulness of features by measuring their impact on the reducts that contain them. The quality of the developed method was evaluated experimentally on both artificial and real-life datasets.