Complexity Measures for Multi-Objective Symbolic Regression

Created by W.Langdon from gp-bibliography.bib Revision:1.4504

  author =       "Michael Kommenda and Andreas Beham and 
                 Michael Affenzeller and Gabriel K. Kronberger",
  title =        "Complexity Measures for Multi-Objective Symbolic
  booktitle =    "Computer Aided Systems Theory, EUROCAST 2015",
  year =         "2015",
  editor =       "Roberto Moreno-Diaz and Franz Pichler and 
                 Alexis Quesada-Arencibia",
  volume =       "9520",
  series =       "Lecture Notes in Computer Science",
  pages =        "409--416",
  address =      "Las Palmas, Gran Canaria, Spain",
  month =        feb,
  publisher =    "Springer",
  keywords =     "genetic algorithms, genetic programming, Symbolic
                 regression, Complexity measures, Multi-objective
                 optimization, NSGA-II",
  isbn13 =       "978-3-319-27340-2",
  URL =          "",
  DOI =          "doi:10.1007/978-3-319-27340-2_51",
  abstract =     "Multi-objective symbolic regression has the advantage
                 that while the accuracy of the learned models is
                 maximized, the complexity is automatically adapted and
                 need not be specified a-priori. The result of the
                 optimization is not a single solution any more, but a
                 whole Pareto-front describing the trade-off between
                 accuracy and complexity.

                 In this contribution we study which complexity measures
                 are most appropriately used in symbolic regression when
                 performing multi- objective optimization with NSGA-II.
                 Furthermore, we present a novel complexity measure that
                 includes semantic information based on the function
                 symbols occurring in the models and test its effects on
                 several benchmark datasets. Results comparing multiple
                 complexity measures are presented in terms of the
                 achieved accuracy and model length to illustrate how
                 the search direction of the algorithm is affected.",

Genetic Programming entries for Michael Kommenda Andreas Beham Michael Affenzeller Gabriel Kronberger