GAssist - Genetic Classifier System

GAssist is a Pittsburgh-style learning classifier system (LCS). It uses a standard genetic algorithm to evolve a population of individuals, each of them being a complete and variable-length rule set.

This system incorporates several mechanisms to tackle data mining problems: A windowing system Incremental Learning with Alternative Strata (ILAS) to improve its efficiency, a representation for continuous datasets called Adaptive Discretization Intervals (ADI), an explicit default rule mechanism and a fitness function based on the Minimum Description Lenghth (MDL) principle to generate accurate and compact solutions. It is intended to deal with problems that can be solved using very compact rule sets.

Documentation

Read the tutorial to quickly learn how to install and use GAssist.

Download

You can download the C++ source code of GAssist.

Related software

If you want to induce rules over a large dataset we recommend BioHEL, a successor of GAssist.

Project team

• Jaume Bacardit
• Maria Franco
• Natalio Krasnogor

Publications

To cite GAssist please use Bacardit2004.

We would also be happy to list your publications on GAssist on this website, so feel free to contact us.

• DOI BibTeX
  M.A. Franco, N. Krasnogor, J. Bacardit
  GAssist vs. BioHEL: critical assessment of two paradigms of genetics-based machine learning
  in Soft Computing, 17(6):953-981, June 2013

  @ARTICLE{Franco2013,
    title = {GAssist vs. BioHEL: critical assessment of two paradigms of genetics-based machine learning},
    author = {Franco, María A. and Krasnogor, Natalio and Bacardit, Jaume},
    year = 2013,
    doi = {10.1007/s00500-013-1016-8},
    month = jun,
    journal = {Soft Computing},
    volume = {17},
    number = {6},
    pages = {953--981}
  }
  

• DOI eprint BibTeX
  J. Bacardit, N. Krasnogor
  Performance and Efficiency of Memetic Pittsburgh Learning Classifier Systems
  in Evolutionary Computation, 17(3):307-342, September 2009

  @ARTICLE{Bacardit2009c,
    title = {Performance and Efficiency of Memetic Pittsburgh Learning Classifier Systems},
    author = {Bacardit, Jaume and Krasnogor, Natalio},
    year = 2009,
    doi = {10.1162/evco.2009.17.3.307},
    month = sep,
    journal = {Evolutionary Computation},
    volume = {17},
    number = {3},
    pages = {307--342}
  }
  

• DOI BibTeX
  M. Stout, J. Bacardit, J.D. Hirst, R.E. Smith, N. Krasnogor
  Prediction of topological contacts in proteins using learning classifier systems
  in Soft Computing, 13(3):245-258, February 2009

  @ARTICLE{Stout2009,
    title = {Prediction of topological contacts in proteins using learning classifier systems},
    author = {Stout, Michael and Bacardit, Jaume and Hirst, Jonathan D. and Smith, Robert E. and Krasnogor, Natalio},
    year = 2009,
    doi = {10.1007/s00500-008-0318-8},
    month = feb,
    journal = {Soft Computing},
    volume = {13},
    number = {3},
    pages = {245--258}
  }
  

• DOI BibTeX
  J. Bacardit, M. Stout, J.D. Hirst, N. Krasnogor
  Data Mining in Proteomics with Learning Classifier Systems
  in Learning Classifier Systems in Data Mining, Studies in Computational Intelligence 125, p.17-46, 2008

  @INCOLLECTION{Bacardit2008a,
    title = {Data Mining in Proteomics with Learning Classifier Systems},
    author = {Bacardit, Jaume and Stout, Michael and Hirst, Jonathan D. and Krasnogor, Natalio},
    year = 2008,
    doi = {10.1007/978-3-540-78979-6_2},
    booktitle = {Learning Classifier Systems in Data Mining},
    publisher = {Springer Berlin Heidelberg},
    volume = {125},
    series = {Studies in Computational Intelligence},
    pages = {17--46}
  }
  

• DOI eprint BibTeX
  J. Bacardit, N. Krasnogor
  Empirical Evaluation of Ensemble Techniques for a Pittsburgh Learning Classifier System
  in Learning Classifier Systems, Lecture Notes in Computer Science 4998, p.255-268, 2008

  @INCOLLECTION{Bacardit2008b,
    title = {Empirical Evaluation of Ensemble Techniques for a Pittsburgh Learning Classifier System},
    author = {Bacardit, Jaume and Krasnogor, Natalio},
    year = 2008,
    doi = {10.1007/978-3-540-88138-4_15},
    booktitle = {Learning Classifier Systems},
    publisher = {Springer Berlin Heidelberg},
    volume = {4998},
    series = {Lecture Notes in Computer Science},
    pages = {255--268}
  }
  

• DOI BibTeX
  M. Tabacman, N. Krasnogor, J. Bacardit, I. Loiseau
  Learning classifier systems for optimisation problems
  in Proceedings of the 2008 GECCO conference companion on Genetic and evolutionary computation - GECCO '08, p.2039, Atlanta, GA, USA, 2008

  @INPROCEEDINGS{Tabacman2008,
    title = {Learning classifier systems for optimisation problems},
    author = {Tabacman, Maximiliano and Krasnogor, Natalio and Bacardit, Jaume and Loiseau, Irene},
    year = 2008,
    doi = {10.1145/1388969.1389018},
    booktitle = {Proceedings of the 2008 GECCO conference companion on Genetic and evolutionary computation - GECCO '08},
    pages = {2039},
    address = {Atlanta, GA, USA}
  }
  

• DOI eprint BibTeX
  J. Bacardit, M. Stout, J.D. Hirst, K. Sastry, X. Llorà, N. Krasnogor
  Automated alphabet reduction method with evolutionary algorithms for protein structure prediction
  in Proceedings of the 9th annual conference on Genetic and evolutionary computation - GECCO '07, p.346, London, England, 2007

  @INPROCEEDINGS{Bacardit2007,
    title = {Automated alphabet reduction method with evolutionary algorithms for protein structure prediction},
    author = {Bacardit, Jaume and Stout, Michael and Hirst, Jonathan D. and Sastry, Kumara and Llorà, Xavier and Krasnogor, Natalio},
    year = 2007,
    doi = {10.1145/1276958.1277033},
    booktitle = {Proceedings of the 9th annual conference on Genetic and evolutionary computation  - GECCO '07},
    pages = {346},
    address = {London, England}
  }
  

• DOI BibTeX
  J. Bacardit, D.E. Goldberg, M.V. Butz
  Improving the Performance of a Pittsburgh Learning Classifier System Using a Default Rule
  in Learning Classifier Systems, Lecture Notes in Computer Science 4399, p.291-307, 2007

  @INCOLLECTION{Bacardit2007a,
    title = {Improving the Performance of a Pittsburgh Learning Classifier System Using a Default Rule},
    author = {Bacardit, Jaume and Goldberg, David E. and Butz, Martin V.},
    year = 2007,
    doi = {10.1007/978-3-540-71231-2_20},
    booktitle = {Learning Classifier Systems},
    publisher = {Springer Berlin Heidelberg},
    volume = {4399},
    series = {Lecture Notes in Computer Science},
    pages = {291--307}
  }
  

• DOI BibTeX
  J. Bacardit, M.V. Butz
  Data Mining in Learning Classifier Systems: Comparing XCS with GAssist
  in Learning Classifier Systems, Lecture Notes in Computer Science 4399, p.282-290, 2007

  @INCOLLECTION{Bacardit2007b,
    title = {Data Mining in Learning Classifier Systems: Comparing XCS with GAssist},
    author = {Bacardit, Jaume and Butz, Martin V.},
    year = 2007,
    doi = {10.1007/978-3-540-71231-2_19},
    booktitle = {Learning Classifier Systems},
    publisher = {Springer Berlin Heidelberg},
    volume = {4399},
    series = {Lecture Notes in Computer Science},
    pages = {282--290}
  }
  

• DOI BibTeX
  J. Bacardit, J.M. Garrell
  Bloat Control and Generalization Pressure Using the Minimum Description Length Principle for a Pittsburgh Approach Learning Classifier System
  in Learning Classifier Systems, Lecture Notes in Computer Science 4399, p.59-79, 2007

  @INCOLLECTION{Bacardit2007c,
    title = {Bloat Control and Generalization Pressure Using the Minimum Description Length Principle for a Pittsburgh Approach Learning Classifier System},
    author = {Bacardit, Jaume and Garrell, Josep Maria},
    year = 2007,
    doi = {10.1007/978-3-540-71231-2_5},
    booktitle = {Learning Classifier Systems},
    publisher = {Springer Berlin Heidelberg},
    volume = {4399},
    series = {Lecture Notes in Computer Science},
    pages = {59--79}
  }
  

• DOI BibTeX
  J. Bacardit, M. Stout, N. Krasnogor, J.D. Hirst, J. Blazewicz
  Coordination number prediction using learning classifier systems
  in Proceedings of the 8th annual conference on Genetic and evolutionary computation - GECCO '06, p.247, Seattle, Washington, USA, 2006

  @INPROCEEDINGS{Bacardit2006,
    title = {Coordination number prediction using learning classifier systems},
    author = {Bacardit, Jaume and Stout, Michael and Krasnogor, Natalio and Hirst, Jonathan D. and Blazewicz, Jacek},
    year = 2006,
    doi = {10.1145/1143997.1144041},
    booktitle = {Proceedings of the 8th annual conference on Genetic and evolutionary computation  - GECCO '06},
    pages = {247},
    address = {Seattle, Washington, USA}
  }
  

• DOI BibTeX
  J. Bacardit, N. Krasnogor
  Smart crossover operator with multiple parents for a Pittsburgh learning classifier system
  in Proceedings of the 8th annual conference on Genetic and evolutionary computation - GECCO '06, p.1441, Seattle, Washington, USA, 2006

  @INPROCEEDINGS{Bacardit2006a,
    title = {Smart crossover operator with multiple parents for a Pittsburgh learning classifier system},
    author = {Bacardit, Jaume and Krasnogor, Natalio},
    year = 2006,
    doi = {10.1145/1143997.1144235},
    booktitle = {Proceedings of the 8th annual conference on Genetic and evolutionary computation  - GECCO '06},
    pages = {1441},
    address = {Seattle, Washington, USA}
  }
  

• DOI BibTeX
  M. Stout, J. Bacardit, J.D. Hirst, N. Krasnogor, J. Blazewicz
  From HP Lattice Models to Real Proteins: Coordination Number Prediction Using Learning Classifier Systems
  in Applications of Evolutionary Computing, Lecture Notes in Computer Science 3907, p.208-220, 2006

  @INCOLLECTION{Stout2006a,
    title = {From HP Lattice Models to Real Proteins: Coordination Number Prediction Using Learning Classifier Systems},
    author = {Stout, Michael and Bacardit, Jaume and Hirst, Jonathan D. and Krasnogor, Natalio and Blazewicz, Jacek},
    year = 2006,
    doi = {10.1007/11732242_19},
    booktitle = {Applications of Evolutionary Computing},
    publisher = {Springer Berlin Heidelberg},
    volume = {3907},
    series = {Lecture Notes in Computer Science},
    pages = {208--220}
  }
  

• DOI BibTeX
  J. Bacardit
  Analysis of the initialization stage of a Pittsburgh approach learning classifier system
  in Proceedings of the 2005 conference on Genetic and evolutionary computation - GECCO '05, p.1843, Washington DC, USA, 2005

  @INPROCEEDINGS{Bacardit2005,
    title = {Analysis of the initialization stage of a Pittsburgh approach learning classifier system},
    author = {Bacardit, Jaume},
    year = 2005,
    doi = {10.1145/1068009.1068321},
    booktitle = {Proceedings of the 2005 conference on Genetic and evolutionary computation  - GECCO '05},
    pages = {1843},
    address = {Washington DC, USA}
  }
  

• eprint BibTeX
  J. Bacardit
  Pittsburgh Genetics-Based Machine Learning in the Data Mining era: Representations, generalization, and run-time
  PhD thesis, Ramon Llull University, Barcelona, Spain, 2004

  @PHDTHESIS{Bacardit2004,
    title = {Pittsburgh Genetics-Based Machine Learning in the Data Mining era: Representations, generalization, and run-time},
    author = {Bacardit, Jaume},
    year = 2004,
    school = {Ramon Llull University, Barcelona, Spain}
  }