Geodesic
Neural network realizations of Bayes decision rules for exponentially distributed data
Kybernetika, Tome 34 (1998) no. 5, pp. 497-514 Cet article a éte moissonné depuis la source Czech Digital Mathematics Library

Voir la notice de l'article

For general Bayes decision rules there are considered perceptron approximations based on sufficient statistics inputs. A particular attention is paid to Bayes discrimination and classification. In the case of exponentially distributed data with known model it is shown that a perceptron with one hidden layer is sufficient and the learning is restricted to synaptic weights of the output neuron. If only the dimension of the exponential model is known, then the number of hidden layers will increase by one and also the synaptic weights of neurons from both hidden layers have to be learned.
For general Bayes decision rules there are considered perceptron approximations based on sufficient statistics inputs. A particular attention is paid to Bayes discrimination and classification. In the case of exponentially distributed data with known model it is shown that a perceptron with one hidden layer is sufficient and the learning is restricted to synaptic weights of the output neuron. If only the dimension of the exponential model is known, then the number of hidden layers will increase by one and also the synaptic weights of neurons from both hidden layers have to be learned.
Classification : 62C10, 62H30, 62M45, 68T05
Keywords: exponentially distributed data 
@article{KYB_1998_34_5_a0,
     author = {Vajda, Igor and Lonek, Belom{\'\i}r and Nikolov, Viktor and Vesel\'y, Arno\v{s}t},
     title = {Neural network realizations of {Bayes} decision rules for exponentially distributed data},
     journal = {Kybernetika},
     pages = {497--514},
     year = {1998},
     volume = {34},
     number = {5},
     mrnumber = {1663720},
     zbl = {1274.62645},
     language = {en},
     url = {http://geodesic.mathdoc.fr/item/KYB_1998_34_5_a0/}
}
TY  - JOUR
AU  - Vajda, Igor
AU  - Lonek, Belomír
AU  - Nikolov, Viktor
AU  - Veselý, Arnošt
TI  - Neural network realizations of Bayes decision rules for exponentially distributed data
JO  - Kybernetika
PY  - 1998
SP  - 497
EP  - 514
VL  - 34
IS  - 5
UR  - http://geodesic.mathdoc.fr/item/KYB_1998_34_5_a0/
LA  - en
ID  - KYB_1998_34_5_a0
ER  - 
%0 Journal Article
%A Vajda, Igor
%A Lonek, Belomír
%A Nikolov, Viktor
%A Veselý, Arnošt
%T Neural network realizations of Bayes decision rules for exponentially distributed data
%J Kybernetika
%D 1998
%P 497-514
%V 34
%N 5
%U http://geodesic.mathdoc.fr/item/KYB_1998_34_5_a0/
%G en
%F KYB_1998_34_5_a0
Vajda, Igor; Lonek, Belomír; Nikolov, Viktor; Veselý, Arnošt. Neural network realizations of Bayes decision rules for exponentially distributed data. Kybernetika, Tome 34 (1998) no. 5, pp. 497-514. http://geodesic.mathdoc.fr/item/KYB_1998_34_5_a0/

[1] Berger J. O.: Statistical Decision Theory and Bayesian Analysis. Second edition. Springer, New York 1985 | MR | Zbl

[2] Brown L. D.: Fundamentals of Statistical Exponential Families. Lecture Notes 9. Inst. of Mathem. Statist., Hayward, California 1986 | MR | Zbl

[3] Bock H. H.: A clustering technique for maximizing $\phi $-divergence, noncentrality and discriminating power. In: Analyzing and Modelling Data and Knowledge (M. Schader, ed.), Springer, Berlin 1992, pp. 19–36

[4] Devijver P., Kittler J.: Pattern Recognition: A Statistical Approach. Prentice Hall, Englewood Cliffs 1982 | MR | Zbl

[5] Funahashi K.: On the approximate realization of continuous mappings by neural networks. Neural Networks 2 (1989), 183–192 | DOI

[6] Hampel F. R., Rousseeuw P. J., Ronchetti E. M., Stahel W. A.: Robust Statistics: The Approach Based on Influence Functions. Wiley, New York 1986 | MR | Zbl

[7] Hand D. J.: Discrimination and Classification. Wiley, New York 1981 | MR | Zbl

[8] Hornik K., Stinchcombe M., White H.: Multilayer feedforward networks and universal approximation. Neural Networks 2 (1989), 359–366 | DOI

[9] Küchler U., Sørensen M.: Exponential families of stochastic processes: A unifying semimartingale approach. Internat. Statist. Rev. 57 (1989), 123–144 | DOI

[10] Lapedes A. S., Farber R. H.: How neural networks work. In: Evolution, Learning and Cognition (Y. S. Lee, ed.), World Scientific, Singapore 1988, pp. 331–340 | MR

[11] Mood A. M., Graybill F. A., Boes D. C.: Introduction to the Theory of Statistics. Third edition. McGraw–Hill, New York 1974 | Zbl

[12] Müller B., Reinhard J., Strickland M. T.: Neural Networks. Second edition. Springer, Berlin 1995

[13] Ripley B. D.: Statistical aspects of neural networks. In: Networks and Chaos (O. E. Barndorff–Nielsen, J. L. Jensen and W. S. Kendall, eds.), Chapman and Hall, London 1993. pp. 40–123 | MR | Zbl

[14] Vajda I.: About perceptron realizations of Bayesian decisions about random processes. In: IEEE International Conference on Neural Networks, vol. 1, IEEE, 1996, pp. 253–257