Symbolic Regression of Conditional Target Expressions

Created by W.Langdon from gp-bibliography.bib Revision:1.4524

  author =       "Michael F. Korns",
  title =        "Symbolic Regression of Conditional Target
  booktitle =    "Genetic Programming Theory and Practice {VII}",
  year =         "2009",
  editor =       "Rick L. Riolo and Una-May O'Reilly and 
                 Trent McConaghy",
  series =       "Genetic and Evolutionary Computation",
  address =      "Ann Arbor",
  month =        "14-16 " # may,
  publisher =    "Springer",
  chapter =      "13",
  pages =        "211--228",
  keywords =     "genetic algorithms, genetic programming, Abstract
                 Expression Grammars, Differential Evolution, Particle
                 Swarm optimization, DE, PSO, Symbolic Regression",
  isbn13 =       "978-1-4419-1653-2",
  DOI =          "doi:10.1007/978-1-4419-1626-6_13",
  abstract =     "This chapter examines techniques for improving
                 symbolic regression systems in cases where the target
                 expression contains conditionals. In three previous
                 papers we experimented with combining high performance
                 techniques from the literature to produce a large
                 scale, industrial strength, symbolic
                 regression-classification system. Performance metrics
                 across multiple problems show deterioration in accuracy
                 for problems where the target expression contains
                 conditionals. The techniques described herein are shown
                 to improve accuracy on such conditional problems. Nine
                 base test cases, from the literature, are used to test
                 the improvement in accuracy. A previously published
                 regression system combining standard genetic
                 programming with abstract expression grammars, particle
                 swarm optimisation, differential evolution, context
                 aware crossover and age-layered populations is tested
                 on the nine base test cases. The regression system is
                 enhanced with these additional techniques: pessimal
                 vertical slicing, splicing of uncorrelated champions
                 via abstract conditional expressions, and abstract
                 mutation and crossover. The enhanced symbolic
                 regression system is applied to the nine base test
                 cases and an improvement in accuracy is observed.",
  notes =        "part of \cite{Riolo:2009:GPTP}",

Genetic Programming entries for Michael Korns