IJMLC 2016 Vol.6(5): 235-240 ISSN: 2010-3700
DOI: 10.18178/ijmlc.2016.6.5.604

Supervised Learning Methods of Bilinear Neural Network Systems Using Discrete Data

Toshio Ito

Abstract—This paper presents supervised learning methods of neural networks called bilinear neural networks with time delay (BNN). A BNN system was proposed to analyze a weak nonlinear model. In this paper, we propose supervised learning methods of BNN systems using discrete data and continuous curves of the data obtained by curve fitting. We introduce a method for fitting a finite Fourier series to discrete data and show that the fitted curve can be created as the output from a BNN system. By using this fitting method, we propose a method for determining the optimal values for the coefficients of all connections for each neuron in BNN systems.

Index Terms—Supervised learning methods, neural networks, nonlinear model, discrete data, discrete Fourier transform, Fourier series, curve fitting.

T. Ito is with Fujitsu Laboratories Ltd., Kawasaki, 211-8588 Japan (e-mail: ito.toshio@jp.fujitsu.com).

[PDF]

Cite: Toshio Ito, "Supervised Learning Methods of Bilinear Neural Network Systems Using Discrete Data," International Journal of Machine Learning and Computing vol. 6, no. 5, pp. 235-240, 2016.

General Information

  • ISSN: 2010-3700 (Online)
  • Abbreviated Title: Int. J. Mach. Learn. Comput.
  • Frequency: Bimonthly
  • DOI: 10.18178/IJMLC
  • Editor-in-Chief: Dr. Lin Huang
  • Executive Editor:  Ms. Cherry L. Chen
  • Abstracing/Indexing: Scopus (since 2017), Inspec (IET), Google Scholar, Crossref, ProQuest, Electronic Journals Library.
  • E-mail: ijmlc@ejournal.net