Enhanced symbolic regression through local variable transformations

Jiří Kubalík; Erik Derner; Robert Babuška

doi:10.5220/0006505200910100

Enhanced symbolic regression through local variable transformations

Jiří Kubalík, Erik Derner, Robert Babuška

Learning & Autonomous Control

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

10 Citations (Scopus)

Abstract

Genetic programming (GP) is a technique widely used in a range of symbolic regression problems, in particular when there is no prior knowledge about the symbolic function sought. In this paper, we present a GP extension introducing a new concept of local transformed variables, based on a locally applied affine transformation of the original variables. This approach facilitates finding accurate parsimonious models. We have evaluated the proposed extension in the context of the Single Node Genetic Programming (SNGP) algorithm on synthetic as well as real-problem datasets. The results confirm our hypothesis that the transformed variables significantly improve the performance of the standard SNGP algorithm.

Original language	English
Title of host publication	Proceedings of the 9th International Joint Conference on Computational Intelligence (IJCCI 2017)
Publisher	SciTePress
Pages	91-100
Volume	1
ISBN (Print)	978-989-758-274-5
DOIs	https://doi.org/10.5220/0006505200910100
Publication status	Published - 2017
Event	IJCCI 2017: 9th International Joint Conference on Computational Intelligence - Funchal, Madeira, Portugal Duration: 1 Nov 2017 → 3 Nov 2017

Conference

Conference	IJCCI 2017: 9th International Joint Conference on Computational Intelligence
Country/Territory	Portugal
City	Funchal, Madeira
Period	1/11/17 → 3/11/17

Keywords

Data-driven Modeling
Nonlinear Regression
Single Node Genetic Programming
Symbolic Regression

Access to Document

10.5220/0006505200910100

Cite this

@inproceedings{c57aba8b98c845f18ecaede9b35df1bd,

title = "Enhanced symbolic regression through local variable transformations",

abstract = "Genetic programming (GP) is a technique widely used in a range of symbolic regression problems, in particular when there is no prior knowledge about the symbolic function sought. In this paper, we present a GP extension introducing a new concept of local transformed variables, based on a locally applied affine transformation of the original variables. This approach facilitates finding accurate parsimonious models. We have evaluated the proposed extension in the context of the Single Node Genetic Programming (SNGP) algorithm on synthetic as well as real-problem datasets. The results confirm our hypothesis that the transformed variables significantly improve the performance of the standard SNGP algorithm.",

keywords = "Data-driven Modeling, Nonlinear Regression, Single Node Genetic Programming, Symbolic Regression",

author = "Ji{\v r}{\'i} Kubal{\'i}k and Erik Derner and Robert Babu{\v s}ka",

year = "2017",

doi = "10.5220/0006505200910100",

language = "English",

isbn = "978-989-758-274-5",

volume = "1",

pages = "91--100",

booktitle = "Proceedings of the 9th International Joint Conference on Computational Intelligence (IJCCI 2017)",

publisher = "SciTePress",

note = "IJCCI 2017: 9th International Joint Conference on Computational Intelligence ; Conference date: 01-11-2017 Through 03-11-2017",

}

Kubalík, J, Derner, E & Babuška, R 2017, Enhanced symbolic regression through local variable transformations. in Proceedings of the 9th International Joint Conference on Computational Intelligence (IJCCI 2017). vol. 1, SciTePress, pp. 91-100, IJCCI 2017: 9th International Joint Conference on Computational Intelligence, Funchal, Madeira, Portugal, 1/11/17. https://doi.org/10.5220/0006505200910100

Enhanced symbolic regression through local variable transformations. / Kubalík, Jiří; Derner, Erik; Babuška, Robert.
Proceedings of the 9th International Joint Conference on Computational Intelligence (IJCCI 2017). Vol. 1 SciTePress, 2017. p. 91-100.

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Enhanced symbolic regression through local variable transformations

AU - Kubalík, Jiří

AU - Derner, Erik

AU - Babuška, Robert

PY - 2017

Y1 - 2017

N2 - Genetic programming (GP) is a technique widely used in a range of symbolic regression problems, in particular when there is no prior knowledge about the symbolic function sought. In this paper, we present a GP extension introducing a new concept of local transformed variables, based on a locally applied affine transformation of the original variables. This approach facilitates finding accurate parsimonious models. We have evaluated the proposed extension in the context of the Single Node Genetic Programming (SNGP) algorithm on synthetic as well as real-problem datasets. The results confirm our hypothesis that the transformed variables significantly improve the performance of the standard SNGP algorithm.

AB - Genetic programming (GP) is a technique widely used in a range of symbolic regression problems, in particular when there is no prior knowledge about the symbolic function sought. In this paper, we present a GP extension introducing a new concept of local transformed variables, based on a locally applied affine transformation of the original variables. This approach facilitates finding accurate parsimonious models. We have evaluated the proposed extension in the context of the Single Node Genetic Programming (SNGP) algorithm on synthetic as well as real-problem datasets. The results confirm our hypothesis that the transformed variables significantly improve the performance of the standard SNGP algorithm.

KW - Data-driven Modeling

KW - Nonlinear Regression

KW - Single Node Genetic Programming

KW - Symbolic Regression

UR - http://www.scopus.com/inward/record.url?scp=85055275064&partnerID=8YFLogxK

U2 - 10.5220/0006505200910100

DO - 10.5220/0006505200910100

M3 - Conference contribution

AN - SCOPUS:85055275064

SN - 978-989-758-274-5

VL - 1

SP - 91

EP - 100

BT - Proceedings of the 9th International Joint Conference on Computational Intelligence (IJCCI 2017)

PB - SciTePress

T2 - IJCCI 2017: 9th International Joint Conference on Computational Intelligence

Y2 - 1 November 2017 through 3 November 2017

ER -

Enhanced symbolic regression through local variable transformations

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this