Embedding Monte Carlo Search of Features in Tree-Based Ensemble Methods

Created by W.Langdon from gp-bibliography.bib Revision:1.3973

@InProceedings{DBLP:conf/pkdd/MaesGW12,
  author =       "Francis Maes and Pierre Geurts and Louis Wehenkel",
  title =        "Embedding Monte Carlo Search of Features in Tree-Based
                 Ensemble Methods",
  booktitle =    "European Conference on Machine Learning and Knowledge
                 Discovery in Databases, ECML PKDD 2012, Part I",
  year =         "2012",
  editor =       "Peter A. Flach and Tijl {De Bie} and 
                 Nello Cristianini",
  pages =        "191--206",
  volume =       "7523",
  series =       "Lecture Notes in Computer Science",
  bibsource =    "DBLP, http://dblp.uni-trier.de",
  address =      "Bristol UK",
  month =        "24-28 " # sep,
  publisher =    "Springer",
  keywords =     "genetic algorithms, genetic programming, MCTS,
                 Dimensionality Reduction, Feature Selection and
                 Extraction, Embedded Feature Generation, Monte Carlo
                 Search, Decision Trees, Random Forests, Tree Boosting",
  isbn13 =       "978-3-642-33459-7",
  DOI =          "doi:10.1007/978-3-642-33460-3_18",
  size =         "16 pages",
  abstract =     "Feature generation is the problem of automatically
                 constructing good features for a given target learning
                 problem. While most feature generation algorithms
                 belong either to the filter or to the wrapper approach,
                 this paper focuses on embedded feature generation. We
                 propose a general scheme to embed feature generation in
                 a wide range of tree-based learning algorithms,
                 including single decision trees, random forests and
                 tree boosting. It is based on the formalisation of
                 feature construction as a sequential decision making
                 problem addressed by a tractable Monte Carlo search
                 algorithm coupled with node splitting. This leads to
                 fast algorithms that are applicable to large-scale
                 problems. We empirically analyse the performances of
                 these tree-based learners combined or not with the
                 feature generation capability on several standard
                 datasets.",
  notes =        "

                 Feature engineering - the process of identifying a good
                 set of features for a given learning task.

                 reverse polish notation RPN grammar postfix. vowel,
                 segment, spambase, satellite, pendigits, dig44.
                 Baseline, random, step, look-ahead. tree stumps
                 variance.

                 www.ecmlpkdd2012.net/files/2012/09/ECMLPKDD2012booklet.pdf",
}

Genetic Programming entries for Francis Maes Pierre Geurts Louis Wehenkel

Citations