Feature Extraction Using Coevolutionary Genetic
Programming
M. Kotani, and D. Kato. Proceedings of the 2004 IEEE Congress on Evolutionary
Computation, page 614--619. Portland, Oregon, IEEE Press, (20-23 June 2004)
Abstract
We propose a new feature extraction method using
evolutionary computations. The extracted features are
defined as polynomial expressions, which are composed
of the original input pattern. These polynomial
expressions are searched by coevolutionary genetic
programming. We introduce a new fitness function based
on competition between individuals. Experiments are
performed for some databases of UCI repository using
the proposed method and k-Nearest Neighbor rule.
Experimental results show that the proposed method can
preserve the diversity of populations and improve
recognition accuracy on most databases.
%0 Conference Paper
%1 kotani:2004:feucgp
%A Kotani, Manabu
%A Kato, Daisuke
%B Proceedings of the 2004 IEEE Congress on Evolutionary
Computation
%C Portland, Oregon
%D 2004
%I IEEE Press
%K Coevolution Combinatorial Poster Session \& algorithms, behavior, collective genetic numerical optimization, programming,
%P 614--619
%T Feature Extraction Using Coevolutionary Genetic
Programming
%X We propose a new feature extraction method using
evolutionary computations. The extracted features are
defined as polynomial expressions, which are composed
of the original input pattern. These polynomial
expressions are searched by coevolutionary genetic
programming. We introduce a new fitness function based
on competition between individuals. Experiments are
performed for some databases of UCI repository using
the proposed method and k-Nearest Neighbor rule.
Experimental results show that the proposed method can
preserve the diversity of populations and improve
recognition accuracy on most databases.
%@ 0-7803-8515-2
@inproceedings{kotani:2004:feucgp,
abstract = {We propose a new feature extraction method using
evolutionary computations. The extracted features are
defined as polynomial expressions, which are composed
of the original input pattern. These polynomial
expressions are searched by coevolutionary genetic
programming. We introduce a new fitness function based
on competition between individuals. Experiments are
performed for some databases of UCI repository using
the proposed method and k-Nearest Neighbor rule.
Experimental results show that the proposed method can
preserve the diversity of populations and improve
recognition accuracy on most databases.},
added-at = {2008-06-19T17:35:00.000+0200},
address = {Portland, Oregon},
author = {Kotani, Manabu and Kato, Daisuke},
biburl = {https://www.bibsonomy.org/bibtex/235316fda11b4ccd1821e558bb4fa810c/brazovayeye},
booktitle = {Proceedings of the 2004 IEEE Congress on Evolutionary
Computation},
interhash = {dce2cedff8a5e4461b435e8be138ffdd},
intrahash = {35316fda11b4ccd1821e558bb4fa810c},
isbn = {0-7803-8515-2},
keywords = {Coevolution Combinatorial Poster Session \& algorithms, behavior, collective genetic numerical optimization, programming,},
month = {20-23 June},
notes = {CEC 2004 - A joint meeting of the IEEE, the EPS, and
the IEE.},
pages = {614--619},
publisher = {IEEE Press},
timestamp = {2008-06-19T17:43:42.000+0200},
title = {Feature Extraction Using Coevolutionary Genetic
Programming},
year = 2004
}