Genetic programming for non-linear equation fitting to chaotic data

Created by W.Langdon from gp-bibliography.bib Revision:1.3872

@InCollection{oakley:1997:HECcd,
  author =       "E. Howard N. Oakley",
  title =        "Genetic programming for non-linear equation fitting to
                 chaotic data",
  booktitle =    "Handbook of Evolutionary Computation",
  publisher =    "Oxford University Press",
  publisher_2 =  "Institute of Physics Publishing",
  year =         "1997",
  editor =       "Thomas Baeck and David B. Fogel and 
                 Zbigniew Michalewicz",
  chapter =      "section G4.3",
  keywords =     "genetic algorithms, genetic programming",
  ISBN =         "0-7503-0392-1",
  URL =          "http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf",
  DOI =          "doi:10.1201/9781420050387.ptg",
  size =         "5 pages",
  abstract =     "Current techniques for the investigation of chaos in
                 data series require long, noise-free experimental
                 measurements which are seldom available in biological
                 and medical work. Genetic programming was seen to offer
                 potential in a number of ways, and was therefore
                 initially used to forecast future data values from very
                 short and noisy input data. Genetic programming proved
                 to be as effective a forecasting tool as others
                 advocated in the literature. Forecasting error
                 initially increased quickly with increasing length of
                 prediction, then increased more slowly, according to a
                 biphasic pattern described previously; the gradients of
                 each limb may be used as a crude indicator of the sum
                 of positive Lyapunov exponents. Although no
                 S-expression ever exactly replicated that used to
                 generate the data, fittest S-expressions did yield
                 useful structural data. Furthermore, the efficacy of
                 forecasting remained high even when noise was added to
                 the data series. The application of genetic programming
                 to original and surrogate data series may be a useful
                 test between chaos and randomness. Runs on surrogate
                 series failed to achieve the high fitness values seen
                 with real data, and were distinguished by shallow and
                 homogeneous populations of S-expressions.",
  notes =        "Mackey and Glass, 1977",
}

Genetic Programming entries for Howard Oakley

Citations