Multistep Fuzzy Classifier Design with Self-tuning Coevolutionary Algorithm

Roman Sergienko, Eugene Semenkin

2013

Abstract

A method of Michigan and Pittsburgh approaches combining for fuzzy classifier design with evolutionary algorithms is presented. Michigan-style stage provides fast search of fuzzy rules with the best grade of certainty values for different classes and smoothing of randomness at initial population forming. Pittsburgh method provides rules subset search with the best performance and predefined number of the rules and doesn’t require a lot of computational power. Besides self-tuning cooperative-competitive coevolutionary algorithm for strategy adaptation is used on Michigan and Pittsburgh stages of fuzzy classifier design. This algorithm solves the problem of genetic algorithm parameters setting automatically. The next result is multistep fuzzy classifier design based on multiple repetition of previous fuzzy classifier design. After each iteration standard deviation of classification performance decreases and classification performance increases. Results of numerical experiments for machine learning problems from UCI repository are presented. Fuzzy classifier design methods comparison with alternative classification methods by performance value demonstrates advantages of the proposed algorithms.

References

  1. Ishibuchi, H., Nakashima, T., and Murata, T. 1999. Performance evaluation of fuzzy classifier systems for multidimensional pattern classification problems. IEEE Trans. on Systems, Man, and Cybernetics, vol. 29, pp. 601-618.
  2. Herrera, F. 2008. Genetic fuzzy systems: taxonomy, current research trends and prospects. Evol. Intel., vol. 1, no. 1, pp. 27-46.
  3. Holland, J. H. and Reitman, J. S. 1978. Cognitive systems based on adaptive algorithms. In D. A. Waterman and F. Hayes-Roth, editors, Pattern-Directed Inference Systems. Academic Press, San Diego, CA.
  4. Smith, S. F. 1980. A Learning System Based On Genetic Adaptive Algorithms. PhD thesis, Department of Computer Science, University of Pittsburgh, Pennsylvania.
  5. Ishibuchi, H., Nakashima T., and Kuroda, T. 2000. A hybrid fuzzy GBML algorithm for designing compact fuzzy rule-based classification systems. Proc. of 9th IEEE International Conference on Fuzzy Systems, (2000), pp. 706-711.
  6. Sergienko, R. B. and Semenkin, E. S. 2010. Competitive cooperation for strategy adaptation in coevolutionary genetic algorithm for constrained optimization. Proc. of 2010 IEEE Congress on Evolutionary Computation, pp. 1626-1631.
  7. Martín-Muñoz, P. and Moreno-Velo. F. J. 2010. FuzzyCN2: an algorithm for extracting fuzzy classification rule lists. Proc. of 2010 IEEE International Conference on Fuzzy Systems, pp. 1783-1789.
  8. Palacios, A. M., Sanchez L., and es Couso. I. 2010. Preprocessing vague imbalanced datasets and its use in genetic fuzzy classifiers. Proc. of 2010 IEEE International Conference on Fuzzy Systems, pp. 2595- 2602.
  9. Nojima, Y., Kaisho, Y., and Ishibuchi, H. 2010. Accuracy improvement of genetic fuzzy rule selection with candidate rule addition and membership tuning. Proc. of 2010 IEEE International Conference on Fuzzy Systems, pp. 527 - 534.
  10. Schapire, R. 2001. The boosting approach to machine learning: an overview. MSRI Workshop on Nonlinear Estimation and Classification, Berkeley, CA.
  11. Breiman, L. Arcing classifiers. 1998. The Annals of Statistics, vol. 26, pp. 801-849 (Mar. 1998).
  12. Ho, T. K. 1998. The random subspace method for constructing decision forests. IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 20, pp.832- 844 (Aug. 1998).
  13. Zhuravlev, J. I. 1998. An algebraic approach to recognition or classifications problems. In Pattern Recognition and Image Analysis, vol. 8, no. 1, pp. 59-100.
  14. Voroncov, V.V. and Kanevsky, D, Y. 2005. Cooperative coevolution algorithm ensemble learning. Tavricheskiy Vestnik Informatiki I Matematiki (Tavria's Gerald of Informatics and Mathematics), pp. 51-66, (Feb. 2005).
Download


Paper Citation


in Harvard Style

Sergienko R. and Semenkin E. (2013). Multistep Fuzzy Classifier Design with Self-tuning Coevolutionary Algorithm . In Proceedings of the 10th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO, ISBN 978-989-8565-70-9, pages 113-120. DOI: 10.5220/0004426501130120


in Bibtex Style

@conference{icinco13,
author={Roman Sergienko and Eugene Semenkin},
title={Multistep Fuzzy Classifier Design with Self-tuning Coevolutionary Algorithm},
booktitle={Proceedings of the 10th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,},
year={2013},
pages={113-120},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0004426501130120},
isbn={978-989-8565-70-9},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 10th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,
TI - Multistep Fuzzy Classifier Design with Self-tuning Coevolutionary Algorithm
SN - 978-989-8565-70-9
AU - Sergienko R.
AU - Semenkin E.
PY - 2013
SP - 113
EP - 120
DO - 10.5220/0004426501130120