@inproceedings{c34cf178401644d68151f8a6495c4959,
title = "Improving RBF networks by the feature selection approach EUBAFES",
abstract = "The curse of dimensionality is one of the severest problems concerning the application of RBF networks. The number of RBF nodes and therefore the number of training examples needed grows exponentially with the intrinsic dimensionality of the input space. One way to address this problem is the application of feature selection as a data preprocessing step. In this paper we propose a two-step approach for the determination of an optimal feature subset: First, all possible feature-subsets are reduced to those with best discrimination properties by the application of the fast and robust filter technique EUBAFES. Secondly we use a wrapper approach to judge, which of the pre-selected feature subsets leads to RBF networks with least complexity and best classification accuracy. Experiments are undertaken to show the improvement for RBF networks by our feature selection approach.",
author = "M. Scherf and W. Brauer",
note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 1997.; 7th International Conference on Artificial Neural Networks, ICANN 1997 ; Conference date: 08-10-1997 Through 10-10-1997",
year = "1997",
doi = "10.1007/bfb0020186",
language = "English",
isbn = "3540636315",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "391--396",
editor = "Wulfram Gerstner and Alain Germond and Martin Hasler and Jean-Daniel Nicoud",
booktitle = "Artificial Neural Networks - ICANN 1997 - 7th International Conference, Proceeedings",
}