EUVIMEDEUVIMED entdecken
Crossref

Improving Model-Based Genetic Programming for Symbolic Regression of Small Expressions

M Virgolin, T Alderliesten, C Witteveen, P A N Bosman

Evolutionary Computation · 01.06.2021 · 29 (2) · 211-237

DOI: 10.1162/evco_a_00278 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Evolutionary Computation, online erschienen am 01.06.2021. Verfasst von M Virgolin und 3 weiteren Autor:innen. Beteiligte Einrichtung: Life Science and Health group, Centrum Wiskunde & Informatica, Amsterdam, 1098 XG, the Netherlands marco.virgolin@cwi.nl. Hauptthemen laut MEDLINE: Algorithms, Biological Evolution. Die Arbeit wurde bisher 131-mal zitiert (Crossref) und nennt 61 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The Gene-pool Optimal Mixing Evolutionary Algorithm (GOMEA) is a model-based EA framework that has been shown to perform well in several domains, including Genetic Programming (GP). Differently from traditional EAs where variation acts blindly, GOMEA learns a model of interdependencies within the genotype, that is, the linkage, to estimate what patterns to propagate. In this article, we study the role of Linkage Learning (LL) performed by GOMEA in Symbolic Regression (SR). We show that the non-uniformity in the distribution of the genotype in GP populations negatively biases LL, and propose a method to correct for this. We also propose approaches to improve LL when ephemeral random constants are used. Furthermore, we adapt a scheme of interleaving runs to alleviate the burden of tuning the population size, a crucial parameter for LL, to SR. We run experiments on 10 real-world datasets, enforcing a strict limitation on solution size, to enable interpretability. We find that the new LL method outperforms the standard one, and that GOMEA outperforms both traditional and semantic GP. We also find that the small solutions evolved by GOMEA are competitive with tuned decision trees, making GOMEA a promising new approach to SR.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

AlgorithmsBiological EvolutionGenetic LinkageSemantics

Stichwörter der Autor:innen

GOMEAGenetic programminginterpretability.linkagemachine learningsymbolic regression

Autor:innen und Institutionen

  • Life Science and Health group, Centrum Wiskunde & Informatica, Amsterdam, 1098 XG, the Netherlands marco.virgolin@cwi.nl.
  • Department of Radiation Oncology, Amsterdam UMC, University of Amsterdam, Amsterdam, 1105 AZ, the Netherlands Department of Radiation Oncology, Leiden University Medical Center, Leiden, 2333 ZA, the Netherlands t.alderl…
  • Algorithmics Group, Delft University of Technology, Delft, 2628 XE, the Netherlands c.witteveen@tudelft.nl.
  • Life Science and Health group, Centrum Wiskunde & Informatica, Amsterdam, 1098 XG, the Netherlands Algorithmics Group, Delft University of Technology, Delft, 2628 XE, the Netherlands peter.bosman@cwi.nl.

Zitieren

APA

Virgolin, M., Alderliesten, T., Witteveen, C., & Bosman, P. A. N. (2021). Improving Model-Based Genetic Programming for Symbolic Regression of Small Expressions. Evolutionary Computation, 29(2), 211-237. https://doi.org/10.1162/evco_a_00278

Vancouver

Virgolin M, Alderliesten T, Witteveen C, Bosman PAN. Improving Model-Based Genetic Programming for Symbolic Regression of Small Expressions. Evolutionary Computation. 2021;29(2):211-237. doi:10.1162/evco_a_00278

BibTeX
@article{Virgolin2021,
  author = {Virgolin, M and Alderliesten, T and Witteveen, C and Bosman, P A N},
  title = {Improving Model-Based Genetic Programming for Symbolic Regression of Small Expressions},
  journal = {Evolutionary Computation},
  year = {2021},
  volume = {29},
  number = {2},
  pages = {211--237},
  doi = {10.1162/evco_a_00278}
}
Kostenlos starten →