Why do multilayer perceptrons have favorable small sample properties?

Machine Intelligence and Pattern Recognition

Volumn 16, Issue C, 1994, Pages 287-298

  Raudys, Šarūnas ^a  

^a VILNIUS UNIVERSITY   (Lithuania)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 85013560142     PISSN: 09230459     EISSN: None     Source Type: Book Series    
DOI: 10.1016/B978-0-444-81892-8.50030-4     Document Type: Chapter

References (24)

1. Jain, A., Chandrasekaran, B., Dimensionality and sample size considerations in pattern recognition practice”. Handbook of Statistics, V. 2, 1982, North Holland, 835–855.
2. Duda, R.O., Hart, P.E., Pattern recognition and scene analysis, 1972, Wiley, New York.
3. Vapnik, V.N., Estimation of dependencies based on empirical data. Moscow, 1979, 1982, Springer Verlag, 448 (in Russian). English translation, 1982.
4. Blumer, A., Ehrenfeucht, A., Hausssler, D., Warmuth, M., Lernability and Vapnik-Chervonenkis dimension. J. Assoc. Comp. Mach. 34:No 4 (1989), 929–965.
5. Haussler D., M.Kearns and R.Schapire. Unifying bounds on the sample complexity of Bayessian learning using information theory and the VC dimension. In Proceedings of the 4th Annual Workshop on Computational Learning Theory, Morgan Kaufmann, San Meteo, CA, pp.61–74., 1991.
6. Duin, R.P.W. Superleaming capabilities of neural networks, In book Proc. of the 8th Scandinavian Conference on Image Analysis NOVIM, Norwegian Society for Image Processing and Pattern Recognition, Tromso, Norway, pp. 547–554, 1993.
7. Kraaijeveld, M.A. Small sample behavior of multi-layer feedforward network classifiers: theoretical and practical aspects–Ph. D. Thesis. Department of Applied Physics, Delft University of Technology, Delft, 1993.
8. Schmidt, W.F. Neural pattern classifying systems, theory and experiments with trainable pattern classifiers–Ph. D. thesis. Department of Applied Physics, Delft University of Technology, Delft, 1993.
9. Raudys Š. Investigation of nonparametric classifiers when training sample size is limited.–Statistical Methods of Control, Issue 14, Inst. of Physics and Math. Press, Vilnius, pp. 117–126, 1974(in Russian).
10. Raudys, Š., On the effectiveness of Parzen window classifier. Int. J. Informatica, Vilnius 2:No 3 (1991), 434–454.
11. Raudys Š. On the amount of a priori information in designing the classification algorithm.–Proc. Acad, of Sciences of the USSR, Technical. Cybernetics., No. 4, pp. 168–174, 1972,(in Russian).
12. Raudys, Š., Pikelis, V., On dimensionality, sample size, classification error and complexity of classification algorithm in pattern recognition. IEEE Trans, on Pattern Analysis and Machine Intelligence PAMI-2:No. 3 (1980), 242–252.
13. Raudys, Š., Jain, A.K., Small sample size effects in statistical pattern recognition: Recommendations for practitioners. IEEE Trans, on Pattern Analysis and Machine Intelligence PAMI-13 (1991), 252–264.
14. Deev, A.D., Discriminant function designed on independent blocks of of variables–Engineering Cybernetics. J. of Academy of Sciences of USSR 12:No 3 (1974), 153–156 (in Russian).
15. Meshalkin, L., Assignment of numerical values to nominal variables.–Statistical Methods of Control, Issue 14, 1974, Inst, of Physics and Math. Press, Vilnius, 149–156 (in Russian).
16. Meshalkin, L.D., Serdobolskij, V.I., Errors in classifying multivariate observations. Theory of Probabilities and Its Applications 23:No. 4 (1978), 772–781 (in Russian).
17. Pikelis, V. Errors of a linear classifier with independent measurements when the learning sample size is small. -Statistical Methods of Control, Issue 5, Inst, of Physics and Math. Press, Vilnius, pp. 69–101, 1973 (in Russian).
18. Baum, E.B. When are k-nearest neighbor and back propagation accurate or feasible sized sets of examples? In (editors L.B. Almeida and C.J. Wellekens) Proceedings of the EURASP Workshop on Neural Networks. Pattern Recognition Group, Department of Applied Physics, Delft University of Technology, Delft, 1990.
19. Raudys, Š., On shape of pattern error function, initializations and intrinsic dimensionality in ANN classifier design. Int. Journal Informatica, Vilnius 4:No 3–4 (1993), 360–383.
20. Raudys Š. Generalization error of maximum margin classifier, In preparation.
21. Boser, B., I.Guyon and V.Vapnik. A training algorithm for optimal margin classifiers.–ACM workshop COLT1992, Pittsburgh, 1992.
22. Vapnik, V.N. A method of minimization of sum risk in problem of pattern recognition.–Statistical Problems of Control, Issue 14, Inst, of Physics and Math. Press, Vilnius, pp. 174–177, 1976 (in Russian).
23. Mao, J. and A.Jain. Regularization techniques in artificial neural networks.–Proc. World Congress on Neural Networks, Portland, July 1993.
24. Raudys Š., M.Skurikhina, T.Cibas and P.Gallinari. Ridge estimates of the covariance matrix and regularization of Artificial Neural Network classifier, Submitted to Pattern Recognition and Image Processing, Int. Journal of Russian Academy of Sciences, Moscow, 1994.