Regularization Approach to Inductive Genetic
Programming
N. Nikolaev, and H. Iba. IEEE Transactions on Evolutionary Computing, 54 (4):
359--375(August 2001)
Abstract
This paper presents an approach to regularization of
inductive genetic programming tuned for learning
polynomials. The objective is to achieve optimal
evolutionary performance when searching high-order
multivariate polynomials represented as tree
structures. We show how to improve the genetic
programming of polynomials by balancing its statistical
bias with its variance. Bias reduction is achieved by
employing a set of basis polynomials in the tree nodes
for better agreement with the examples. Since this
often leads to over-fitting, such tendencies are
counteracted by decreasing the variance through
regularization of the fitness function. We demonstrate
that this balance facilitates the search as well as
enables discovery of parsimonious, accurate, and
predictive polynomials. The experimental results given
show that this regularization approach outperforms
traditional genetic programming on benchmark data
mining and practical time-series prediction tasks.
%0 Journal Article
%1 nikolaev:2001:TEC
%A Nikolaev, Nikolay Y.
%A Iba, Hitoshi
%D 2001
%J IEEE Transactions on Evolutionary Computing
%K (artificial Kolmogorov-Gabor STROGANOFF,local algorithms, bias, data genetic inductive intelligence), learning mining, multivariate nodes, polynomials, prediction, programming, regularization, search searching, series statistical structures, time tree
%N 4
%P 359--375
%T Regularization Approach to Inductive Genetic
Programming
%U http://ieeexplore.ieee.org/iel5/4235/20398/00942530.pdf?isNumber=20398
%V 54
%X This paper presents an approach to regularization of
inductive genetic programming tuned for learning
polynomials. The objective is to achieve optimal
evolutionary performance when searching high-order
multivariate polynomials represented as tree
structures. We show how to improve the genetic
programming of polynomials by balancing its statistical
bias with its variance. Bias reduction is achieved by
employing a set of basis polynomials in the tree nodes
for better agreement with the examples. Since this
often leads to over-fitting, such tendencies are
counteracted by decreasing the variance through
regularization of the fitness function. We demonstrate
that this balance facilitates the search as well as
enables discovery of parsimonious, accurate, and
predictive polynomials. The experimental results given
show that this regularization approach outperforms
traditional genetic programming on benchmark data
mining and practical time-series prediction tasks.
@article{nikolaev:2001:TEC,
abstract = {This paper presents an approach to regularization of
inductive genetic programming tuned for learning
polynomials. The objective is to achieve optimal
evolutionary performance when searching high-order
multivariate polynomials represented as tree
structures. We show how to improve the genetic
programming of polynomials by balancing its statistical
bias with its variance. Bias reduction is achieved by
employing a set of basis polynomials in the tree nodes
for better agreement with the examples. Since this
often leads to over-fitting, such tendencies are
counteracted by decreasing the variance through
regularization of the fitness function. We demonstrate
that this balance facilitates the search as well as
enables discovery of parsimonious, accurate, and
predictive polynomials. The experimental results given
show that this regularization approach outperforms
traditional genetic programming on benchmark data
mining and practical time-series prediction tasks.},
added-at = {2008-06-19T17:46:40.000+0200},
author = {Nikolaev, Nikolay Y. and Iba, Hitoshi},
biburl = {https://www.bibsonomy.org/bibtex/2722b45ceed96e2d92c2c9f8f8f3c1b48/brazovayeye},
interhash = {5751939b9efe0deb309b1891eb790aac},
intrahash = {722b45ceed96e2d92c2c9f8f8f3c1b48},
journal = {IEEE Transactions on Evolutionary Computing},
keywords = {(artificial Kolmogorov-Gabor STROGANOFF,local algorithms, bias, data genetic inductive intelligence), learning mining, multivariate nodes, polynomials, prediction, programming, regularization, search searching, series statistical structures, time tree},
month = {August},
number = 4,
pages = {359--375},
timestamp = {2008-06-19T17:48:24.000+0200},
title = {Regularization Approach to Inductive Genetic
Programming},
url = {http://ieeexplore.ieee.org/iel5/4235/20398/00942530.pdf?isNumber=20398},
volume = 54,
year = 2001
}