by Rafał Scherer, Marcin Korytkowski, Robert Nowicki, Leszek Rutkowski
Abstract:
In the paper we propose a new class of modular systems for classification in the case of missing features. We incorporate the rough set theory into construction of neuro-fuzzy systems which create the modular structure. The AdaBoost algorithm is combined with the gradient algorithm to train the whole system. We illustrate the performance of our approach on typical benchmarks.
Reference:
R. Scherer, M. Korytkowski, R. Nowicki, L. Rutkowski, "Modular Rough Neuro-fuzzy Systems for Classification", Lecture Notes in Computer Science, vol. 4967, 2008, pp. 540-548.
Bibtex Entry:
@ARTICLE{SchKorNowRut_PPAM2008,
author = {Rafał Scherer and Marcin Korytkowski and Robert Nowicki and Leszek
Rutkowski},
title = {Modular Rough Neuro-fuzzy Systems for Classification},
journal = {Lecture Notes in Computer Science},
year = {2008},
volume = {4967},
pages = {540-548},
abstract = {In the paper we propose a new class of modular systems for classification
in the case of missing features. We incorporate the rough set theory
into construction of neuro-fuzzy systems which create the modular
structure. The AdaBoost algorithm is combined with the gradient algorithm
to train the whole system. We illustrate the performance of our approach
on typical benchmarks.},
url = {http://www.springerlink.com/content/372n3787p08x817m/}
}