ML Anthology

Specification of Training Sets and the Number of Hidden Neurons for Multilayer Perceptrons

Laurisete dos Santos Camargo, Takashi Yoneyama
NeCo 2001 pp. 2673-2680
doi:10.1162/089976601317098484 /neco/2001/dossantoscamargo2001neco-specification/

Abstract

This work concerns the selection of input-output pairs for improved training of multilayer perceptrons, in the context of approximation of univariate real functions. A criterion for the choice of the number of neurons in the hidden layer is also provided. The main idea is based on the fact that Chebyshev polynomials can provide approximations to bounded functions up to a prescribed tolerance, and, in turn, a polynomial of a certain order can be fitted with a three-layer perceptron with a prescribed number of hidden neurons. The results are applied to a sensor identification example.

NeCo Semantic Scholar

Cite

Text

dos Santos Camargo and Yoneyama. "Specification of Training Sets and the Number of Hidden Neurons for Multilayer Perceptrons." Neural Computation, 2001. doi:10.1162/089976601317098484

Markdown

[dos Santos Camargo and Yoneyama. "Specification of Training Sets and the Number of Hidden Neurons for Multilayer Perceptrons." Neural Computation, 2001.](https://mlanthology.org/neco/2001/dossantoscamargo2001neco-specification/) doi:10.1162/089976601317098484

BibTeX

@article{dossantoscamargo2001neco-specification,
  title     = {{Specification of Training Sets and the Number of Hidden Neurons for Multilayer Perceptrons}},
  author    = {dos Santos Camargo, Laurisete and Yoneyama, Takashi},
  journal   = {Neural Computation},
  year      = {2001},
  pages     = {2673-2680},
  doi       = {10.1162/089976601317098484},
  volume    = {13},
  url       = {https://mlanthology.org/neco/2001/dossantoscamargo2001neco-specification/}
}