Prog. Theor. Phys. Vol. 91 No. 2 (1994) pp. 397-402

[ Full Text PDF : FREE ACCESS (597K) ]

Letters

Analysis of Learning Processes of Chaotic Time Series by Neural Networks

Tsuyoshi Hondou and Yasuji Sawada

Research Institute of Electrical Communication, Tohoku University, Sendai 980

(Received November 2, 1993)

Abstract:

Study of the learning process of chaotic time series by neural networks with back-propagation algorithm showed the existence of a general learning process. The process was found to be efficiently characterized by newly defined normalized measures, the coherence and the direction-cosine for the motion of weight vectors.

URL : http://ptp.ipap.jp/link?PTP/91/397/
DOI : 10.1143/PTP.91.397

References:

D. E. Rumelhart, J. L. McClelland and the PDP Research Group, Parallel Distributed Processing (MIT Press, 1986).
F. J. Pineda, Phys. Rev. Lett. 59 (1987), 2229[APS].
K. Funahashi, Neural Networks 2 (1989), 183.
K. Doya and S. Yoshizawa, Proceedings of the 3rd IJCNN 1 (1989), p. 27.
M. Sato, Y. Murakami and K. Joe, Proc. Inter. Conf. on Fuzzy Logic and Neural Networks (1990), p. 601.
J. Moody and C. Darken, Neural Comput. 1 (1989), 281.
D. J. Amit, Modeling Brain Function (Cambridge Univ. Press, 1989).
T. M. Heskes, E. T. P. Slijpen and B. Kappen, Phys. Rev. A46 (1992), 5221[APS].
H. Suhwarze, M. Opper and W. Kinzel, Phys. Rev. A46 (1992), 6185[APS].
H. Horner, Z. Phys. B87 (1992), 371.
M. Opper and D. Haussler, Phys. Rev. Lett. 66 (1991), 2677[APS].
H. S. Seung and H. Sompolinsky, Phys. Rev. A45 (1992), 6056[APS].
J. A. McGeoch and A. L. Irion, The Psychology of Human Learning, 2nd ed. (Longmans, New York, 1952).
The data will be described elsewhere.
T. Hondou and Y. Sawada, Proceedings of IJCNN-Nagoya (1993), p. 2387.
H. G. Schuster, Deterministic Chaos (Physik-Verlag, Germany, 1984).
We used r = 1.99 for the simulation, since at r = 2 iteration of f can lead x = 0, which destroys the precision of our calculation.
For example, S. Amari, IEEE Trans. EC-16(3) (1967), 299.

Citing Article(s) :

Journal of the Physical Society of Japan 63 (1994) pp. 2014-2015 :
Symmetry Breaking by Correlated Noise in a Multistable System

Tsuyoshi Hondou
Progress of Theoretical Physics Vol. 93 No. 5 (1995) pp. 845-856 :
Back Propagation Learning Using a Super Stable Motion

Masayoshi Inoue, Akira Tanaka and Kenji Nakamoto
Progress of Theoretical Physics Vol. 95 No. 4 (1996) pp. 817-822 :
Self-Annealing Dynamics in a Multistable System

Tsuyoshi Hondou