Modelling and Prediction Using GMDH Networks
Chapters 2 and 3 have shown that neural networks can be employed to identify dynamic systems. The main advantages of neural networks over conventional identification methods include simplicity of implementation and good approximation properties [Warwick et aI, 1992]. In feedforward network based identification schemes, neural networks are used to represent the implied static mapping between the available input and output data. The network structures (number of layers and number of units in each layer) are predefined and remain unchanged both during and after training. Successful identification is often dependent on proper pre-estimation of the network structure.
KeywordsNeural Network Input Unit Present Layer Good Approximation Property Conventional Identification Method
Unable to display preview. Download preview PDF.
- Barron, RL (1975) Learning networks improve computer-aided prediction and control, Computer Design, 14(8),65–70.Google Scholar
- Duffy, J.J. and Franklin, M.A. (1975) A learning identification algorithm and its application to an environment system, IEEE Trans. Systems, Man, and Cybernetics, 5(2), 226–240.Google Scholar
- Hecht-Nielsen, R. (1990) Neurocomputing, Addison-Wesley Publishing Company.Google Scholar
- Kosko, B. (1992) Neural Networks and Fuzzy system, Englewood Cliffs, NJ: Prentice-Hall.Google Scholar
- Shrier, S., Barron, R.L., Gilstrap, L.O. (1987) Polynomial and neural networks: analogies and engineering applications, Proceedings of IEEE International Conference on Neural Networks, II, 431–439.Google Scholar
- Weisberg, S. (1985) Applied Lineal’ Regression, Wiley, New York, 31–32.Google Scholar