Generalisation Enhancement via Input Space Transformation: A GP Approach

Created by W.Langdon from gp-bibliography.bib Revision:1.4340

  author =       "Ahmed Kattan and Michael Kampouridis and 
                 Alexandros Agapitos",
  title =        "Generalisation Enhancement via Input Space
                 Transformation: A GP Approach",
  booktitle =    "17th European Conference on Genetic Programming",
  year =         "2014",
  editor =       "Miguel Nicolau and Krzysztof Krawiec and 
                 Malcolm I. Heywood and Mauro Castelli and Pablo Garcia-Sanchez and 
                 Juan J. Merelo and Victor M. {Rivas Santos} and 
                 Kevin Sim",
  series =       "LNCS",
  volume =       "8599",
  publisher =    "Springer",
  pages =        "61--74",
  address =      "Granada, Spain",
  month =        "23-25 " # apr,
  organisation = "EvoStar",
  keywords =     "genetic algorithms, genetic programming",
  isbn13 =       "978-3-662-44302-6",
  DOI =          "doi:10.1007/978-3-662-44303-3_6",
  abstract =     "This paper proposes a new approach to improve
                 generalisation of standard regression techniques when
                 there are hundreds or thousands of input variables. The
                 input space X is composed of observational data of the
                 form (x_i, y(x_i)), i = 1... n where each x_i denotes a
                 k-dimensional input vector of design variables and y is
                 the response. Genetic Programming (GP) is used to
                 transform the original input space X into a new input
                 space Z = (z_i, y(z_i)) that has smaller input vector
                 and is easier to be mapped into its corresponding
                 responses. GP is designed to evolve a function that
                 receives the original input vector from each x_i in the
                 original input space as input and return a new vector
                 z_i as an output. Each element in the newly evolved z_i
                 vector is generated from an evolved mathematical
                 formula that extracts statistical features from the
                 original input space. To achieve this, we designed GP
                 trees to produce multiple outputs. Empirical evaluation
                 of 20 different problems revealed that the new approach
                 is able to significantly reduce the dimensionality of
                 the original input space and improve the performance of
                 standard approximation models such as Kriging, Radial
                 Basis Functions Networks, and Linear Regression, and GP
                 (as a regression techniques). In addition, results
                 demonstrate that the new approach is better than
                 standard dimensionality reduction techniques such as
                 Principle Component Analysis (PCA). Moreover, the
                 results show that the proposed approach is able to
                 improve the performance of standard Linear Regression
                 and make it competitive to other stochastic regression
  notes =        "Part of \cite{Nicolau:2014:GP} EuroGP'2014 held in
                 conjunction with EvoCOP2014, EvoBIO2014, EvoMusArt2014
                 and EvoApplications2014",

Genetic Programming entries for Ahmed Kattan Michael Kampouridis Alexandros Agapitos