Selecting Features and Objects for Mixed and Incomplete Data

  • Yenny Villuendas-Rey
  • Milton García-Borroto
  • José Ruiz-Shulcloper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5197)

Abstract

Selecting objects and features before classifying is a very important task, and can lead to big improvements in classifier accuracy and speed. There are many papers about this topic, but few of them consider the simultaneous or combined approach. In this paper, we present a new method for combined object and feature selection for databases with features not purely numeric or non-numeric. The experiments performed show that it attains the best tradeoff between object and feature reduction in 12 of 15 tested databases, without a significant impact in 1-NN accuracy.

Keywords

object selection feature selection supervised classification classifier accuracy mixed and incomplete data 

References

  1. 1.
    Bezdek, J.C., Kuncheva, L.I.: Nearest Prototype classifiers design: an experimental study. Technical Report, University of West Florida, pp. 1–37 (2004)Google Scholar
  2. 2.
    Guyon, I., Elisseeff, A.: An introduction to variable and feature selection. The Journal of Machine Learning Research 3, 1157–1182 (2003)MATHGoogle Scholar
  3. 3.
    Kuncheva, L.I., Jain, L.C.: Nearest neighbor classifier: Simultaneous editing and feature selection. Pattern Recognition Letters 20, 1149–1156 (1999)CrossRefGoogle Scholar
  4. 4.
    Ruiz-Shulcloper, J., Abidi, M.A.: Logical Combinatorial Pattern Recognition: A Review. In: Pandalai, S.G. (ed.) Recent Research Developments in Pattern Recognition. Transworld Research Networks, USA, pp. 133–176 (2002)Google Scholar
  5. 5.
    Santiesteban, Y., Pons-Porrata, A.: LEX: A new algorithm to calculate typical testors. Revista Ciencias Matemáticas 21, 118–126 (2003)Google Scholar
  6. 6.
    García-Borroto, M., Ruiz-Shulcloper, J.: Selecting Prototypes in Mixed Incomplete Data. In: Sanfeliu, A., Cortés, M.L. (eds.) CIARP 2005. LNCS, vol. 3773, pp. 450–459. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  7. 7.
    Skalak, D.B.: Prototype and Feature Selection by Sampling and Random Mutation Hill Climbing Algorithms. In: Eleventh International Machine Learning Conference, pp. 293–301. Morgan Kaufmann, New Brunswick (1994)Google Scholar
  8. 8.
    Ishibushi, H., Nakashima, T.: Evolution of reference sets in nearest neighbor classification. In: McKay, B., Yao, X., Newton, C.S., Kim, J.-H., Furuhashi, T. (eds.) SEAL 1998. LNCS (LNAI), vol. 1585, pp. 82–89. Springer, Heidelberg (1999)CrossRefGoogle Scholar
  9. 9.
    Dasarathy, B.V.: Concurrent Feature and Prototype Selection in the Nearest Neighbor Decision Process. In: 4th World Multiconference on Systemics, Cybernetics and Informatics, vol. VII, pp. 628–633. Orlando, USA (2000)Google Scholar
  10. 10.
    Rozsypal, A., Kubat, M.: Selecting representative examples and attributes by a genetic algorithm. Intelligent Data Analysis 7, 291–304 (2003)MATHGoogle Scholar
  11. 11.
    Villuendas-Rey, Y., García-Borroto, M., Medina-Pérez, M.A., Ruiz-Shulcloper, J.: Simultaneous features and objects selection for Mixed and Incomplete data. In: Martínez-Trinidad, J.F., Carrasco Ochoa, J.A., Kittler, J. (eds.) CIARP 2006. LNCS, vol. 4225, pp. 597–605. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  12. 12.
    Ahn, H., Kim, K.J., Han, I.: A case-based reasoning system with the two-dimensional reduction technique for customer classification. Expert Systems with Applications: An International Journal 32, 1011–1019 (2007)CrossRefGoogle Scholar
  13. 13.
    Wilson, R.D., Martinez, T.R.: Improved Heterogeneous Distance Functions. Journal of Artificial Intelligence Research 6, 1–34 (1997)MathSciNetMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Yenny Villuendas-Rey
    • 1
  • Milton García-Borroto
    • 2
  • José Ruiz-Shulcloper
    • 3
  1. 1.Ciego de Ávila University UNICAC. de ÁvilaCuba
  2. 2.Bioplantas Center, UNICAC. de ÁvilaCuba
  3. 3.Advanced Technologies Applications Center, MINBASCuba

Personalised recommendations