Skip to main content
Log in

Multivariate polynomials: A spanning question

  • Published:
Constructive Approximation Aims and scope

Abstract

The main result of this paper is the following. Ifg is any given polynomial of two variables, then

$$span\left\{ {\left( {g\left( {. - a,. - b} \right)} \right)^k :\left( {a,b} \right) \in R^2 ,k \in {\rm Z}_ + } \right\}$$

contains all polynomials if and only if

$$span\left\{ {g\left( {. - a,. - b} \right):\left( {a,b} \right) \in R^2 } \right\}$$

separates points. This result is not valid inR d ford≥4.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. L. Leshno, V. Ya. Lin, A. Pinkus, S. Schocken (1993):Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Networks,6:861–867.

    Google Scholar 

  2. A. G. Vitushkin, G. M. Henkin (1967):Linear superpositions of functions. Uspekhi Mat. Nauk,22:77–124. (English translation (1967): Russian Math. Surveys,22:77–125.)

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Additional information

Communicated by Tim N. T. Goodman.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Pinkus, A., Wajnryb, B. Multivariate polynomials: A spanning question. Constr. Approx 11, 165–180 (1995). https://doi.org/10.1007/BF01203414

Download citation

  • Received:

  • Revised:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF01203414

AMS classification

Key words and phrases

Navigation