Korean J. Math. Vol. 23 No. 3 (2015) pp.327-336
DOI: https://doi.org/10.11568/kjm.2015.23.3.327

Constructive approximation by neural networks with positive integer weights

Article Sidebar

PDF
Published: Sep 30, 2015
Keywords:
Neural network, Positive integer weight, Sigmoidal function.
Mathematics Subject Classification:
1A25, 41A46.

Main Article Content

Bum Il Hong
Department of Applied Mathematics and The Institute of Natural Sciences Kyung Hee University Yongin 446-701, Korea
Nahmwoo Hahm
Department of Mathematics Incheon National University Incheon 406-772, Korea

Abstract

In this paper, we study a constructive approximation by neural networks with positive integer weights. Like neural networks with real weights, we show that neural networks with positive integer weights can even approximate arbitrarily well for any continuous functions on compact subsets of $\mathbb{R}$. We give a numerical result to justify our theoretical result.


Article Details

Supporting Agencies

This research was supported by Incheon National University Fund 2014.

References

[1] D. Chen, Degree of Approximation by Superpositions of a Sigmoidal Function, Approx. Theory and Appl., 9 (3) (1993), 17–28. Google Scholar

[2] C. K. Chui and X. Li, Approximation by Ridge Functions and Neural Networks with One Hidden Layer, J. Approx. Theory, 70 (1992), 17–28. Google Scholar

[3] N. Hahm and B. I. Hong, Degree of Approximation by Periodic Neural Networks, Korean J. Math., 22 (2) (2014), 307–315. Google Scholar

[4] N. Hahm and B. I. Hong, The Capability of Localized Neural Network Approxi- mation, Honam Math. J., 34 (3) (2012), 341–349. Google Scholar

[5] B. L. Kalman and S. C. Kwasny, Why Tanh : Choosing a Sigmoidal Function, Lect. Notes Comput. Sc., 600 (1995), 303–312. Google Scholar

[6] J. Killian and H. T. Siegelmann, The Dynamic Universality of Sigmoidal Neural Networks, Inform. Comput., 128 (1996), 48–56. Google Scholar

[7] M. Leshno, V. Y. Lin, A. Pinkus and S. Schoken, Multilayer Feedforward Networks with a Nonpolynomial Activation Function Can Approximate Any Func- tion, Neural Networks, 7 (1995), 861–867. Google Scholar

[8] G. Lewicki and G. Marino, Approximation of Functions of Finite Variation by Superpositions of a Sigmoidal Function, Appl. Math. Lett. 17, (2004), 1147– 1152. Google Scholar

[9] M. V. Medvedeva, On Sigmoidal Functions, Mosc. Univ. Math. Bull., 53 (1) (1998), 16–19. Google Scholar