Marcin Blachnik1, Tadeusz Wieczorek1, and
Wlodzislaw Duch2,3,
1Division of Computer Studies,
Department of Electrotechnology,
The Silesian University of Technology, Katowice, Poland;
2School of Computer Engineering,
Nanyang Technological University, Singapore,
3Department of Informatics,
Nicolaus Copernicus University, Grudziadzka 5, 87-100 Torun, Poland.
Abstract.
Probabilistic distance functions, including several variants of value difference metrics, minimum risk metric and Short- Fukunaga metrics, are used with prototype-based rules (P-rules) to provide a very concise and comprehensible classification model. Application of probabilistic metrics to nominal or discrete features is straightforward. Heterogeneous metrics that handle continuous attributes with discretized or interpolated probabilistic metrics were combined with several methods of probability density estimation. Numerical experiments on artificial and real data show the usefulness of such approach as an alternative to neurofuzzy models.
Reference: Blachnik M, Duch W, Wieczorek T, Probabilistic distance measures for prototype-based rules. Proc. of the 12th Int. Conference on Neural Information Processing (ICONIP'2005), Taipei, Taiwan, pp. 445-450
Preprint for comments in PDF, 112 KB.
BACK to the publications of W. Duch.
BACK to the on-line publications of the Department of Informatics, NCU.