An artificial neural networks for approximating polynomial functions

The authors use polynomial function as a common base to measure the capacity of a feedforward artificial neural network (FANN) with a finite number of hidden nodes. They show that there is a relationship between the capacity of a FANN in approximating polynomial functions and the number of hidden nodes used in the FANN. A procedure for realizing a FANN in approximating polynomial functions is described. Two examples are given to show the procedure. Several experiments are reported, verifying that a FANN with a certain number of hidden nodes has the capability to learn a given polynomial function. The experiments also showed that the proposed algorithm for training a FANN performs accurately.<<ETX>>

[1]  Behnam Malakooti,et al.  An adaptive feedforward artificial neural network with the applications to multiple criteria decision making , 1990, 1990 IEEE International Conference on Systems, Man, and Cybernetics Conference Proceedings.

[2]  E. Cheney.  Introduction to approximation theory , 1966 .

[3]  Matthew Zeidenberg,et al.  Neural networks in artificial intelligence , 1990, Ellis Horwood series in artificial intelligence.

[4]  Bernard Widrow,et al.  30 years of adaptive neural networks: perceptron, Madaline, and backpropagation , 1990, Proc. IEEE.

[5]  Yih-Fang Huang,et al.  Bounds on the number of hidden neurons in multilayer perceptrons , 1991, IEEE Trans. Neural Networks.

[6]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[7]  R. L. Keeney,et al.  Decisions with Multiple Objectives: Preferences and Value Trade-Offs , 1977, IEEE Transactions on Systems, Man, and Cybernetics.

[8]  Panos J. Antsaklis,et al.  A simple method to derive bounds on the size and to train multilayer neural networks , 1991, IEEE Trans. Neural Networks.