Authors
Kurt Hornik, Maxwell Stinchcombe, Halbert White
Publication date
1989/1/1
Journal
Neural networks
Volume
2
Issue
5
Pages
359-366
Publisher
Pergamon
Description
This paper rigorously establishes that standard multilayer feedforward networks with as few as one hidden layer using arbitrary squashing functions are capable of approximating any Borel measurable function from one finite dimensional space to another to any desired degree of accuracy, provided sufficiently many hidden units are available. In this sense, multilayer feedforward networks are a class of universal approximators.
Total citations
Scholar articles
K Hornik, M Stinchcombe, H White - Neural networks, 1989