Advertisement

Cybernetics and Systems Analysis

, Volume 48, Issue 4, pp 592–600 | Cite as

Upper-bound estimates for classifiers based on a dissimilarity function

  • B. P. Rusyn
  • V. A. Tayanov
  • O. A. Lutsyka
Article

Abstract

Approaches to calculating upper-bound estimates of recognition probability are proposed that can be used for a more general class of models. One of estimates determines the stability of object coverage by classification algorithms on the basis of distribution of distances between objects, and another estimate is underlain by leave-one-out cross-validation. This considerably simplifies and facilitates the construction of estimates.

Keywords

upper-bound estimate for recognition probability leave-one-out cross-validation (LOOCV) stability of object coverage by classification algorithms distribution of distances right (left) asymmetry metric, class of objects 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    K. V. Vorontsov, “Combinatorial probability and the tightness of generalization bounds,” Pattern Recognition and Image Analysis, 18, No. 2, 243–259 (2008).Google Scholar
  2. 2.
    V. Vapnik, The Nature of Statistical Learning Theory, Springer, New York (2000).MATHGoogle Scholar
  3. 3.
    Yu. I. Zhuravlev, “An algebraic approach to recognition or classification problems,” Problems of Cybernetics, 33, 5–68 (1978).Google Scholar
  4. 4.
    K. V. Vorontsov, Machine Learning and Data Analysis: A Series of Lectures “Mathematical methods of learning from precedents,” http://www.ccas.ru/voron/teaching.html.
  5. 5.
    M. Shlezinger and V. Glavach, Ten Lectures on Statistical and Structural Recognition [in Russian], Naukova Dumka, Kiev (2004).Google Scholar
  6. 6.
    T. K. Moon and W. C. Stirling, Mathematical Methods and Algorithms for Signal Processing, Prentice-Hall, N.Y. (2000).Google Scholar
  7. 7.
    B. E. Kapustii, B. P. Rusyn, and V. A. Tayanov, “Classifier optimization in small sample size condition,” Automatic Control and Computer Sci., 40, No. 5, 17–22 (2006).Google Scholar
  8. 8.
    B. O. Kapustii, B. P. Rusyn, and V. A. Tayanov, “Combinatorial estimation of the influence of a class information coverage reduction on the generalization ability of 1NN classification algorithms,” Artificial Intelligence, No. 1, 49–54 (2008).Google Scholar
  9. 9.
    S. Karlin, Fundamentals of the Theory of Random Processes [Russian translation], Mir, Moscow (1971).Google Scholar
  10. 10.
    V. S. Korolyuk, N. I. Portenko, A. V. Skorokhod, and A. F. Turbin, A Manual on Probability Theory and Mathematical Statistics [in Russian], Nauka, Moscow (1985).Google Scholar
  11. 11.
    C. M. Bishop, Pattern Recognition and Machine Learning (Information Science and Statistics), Springer, London (2006).Google Scholar
  12. 12.
    E. W. Weisstein, Chebyshev inequality, http://mathworld.wolfram.com/Chebyshev Inequality.html, 10.12.2008.
  13. 13.
    E. W. Weisstein, Gauss inequality, http://mathworld.wolfram.com/GaussInequality.html, 10.12.2008.
  14. 14.
    G. Tu. and R. Gonsales, Principles of Pattern Recognition [Russian translation], Mir, Moscow (1978).Google Scholar
  15. 15.
    B. O. Kapustii, B. P. Rusyn, and V. A. Tayanov, “A new approach to the definition of the probability of correct recognition of objects from sets,” USiM, No. 2, 8–13 (2005).Google Scholar

Copyright information

© Springer Science+Business Media, Inc. 2012

Authors and Affiliations

  1. 1.G. V. Karpenko Physico-Mechanical InstituteNational Academy of Sciences of UkraineLvivUkraine

Personalised recommendations