Multilayer feedforward networks are universal approximators

Billingsley, 1979 Cybenko, 1988, Approximation by superpositions of a sigmoidal function Dugundji, 1966 Gallant, 1988, There exists a neural network that does not make avoidable mistables, I:657 Grenander, 1981 Halmos, 1974 Hecht-Nielsen, 1987, Kolmogorov's mapping neural network existence theorem, III:11 Hecht-Nielsen, 1989, Theory of the back propagation neural network, I:593 Hornik, 1988, Multilayer feedforward networks are universal approximators 1987 1988 Irie, 1988, Capabilities of three layer perceptrons, I:641 Kolmogorov, 1957, On the representation of continuous functions of many variables by superposition of continuous functions of one variable and addition, Doklady Akademii Nauk SSR, 114, 953 Kolmogorov, 1961, ϵ-entropy and ϵ-capacity of sets in functional spaces, American Mathematical Society Translations, 2, 277, 10.1090/trans2/017/10 Lapedes, 1988, How neural networks work le Cun, 1987, Modeles connexionistes de l'apprentissage Lorentz, 1976, The thirteenth problem of Hilbert, Vol. 28, 419 Maxwell, 1986, Nonlinear dynamics of artificial neural systems Minsky, 1969 Rudin, 1964 Severini, 1987, Convergence rates of maximum likelihood and related estimates in general parameter spaces Stinchcombe, 1989, Universal approximation using feedforward networks with non-sigmoid hidden layer activation functions, I:613 White, 1988, The case for conceptual and operational separation of network architectures and learning mechanisms White, 1988, Multilayer feedforward networks can learn arbitrary mappings: Connectionist nonparametric regression with automatic and semi-automatic determination of network complexity White, H., & Wooldridge, J. M. (in press). Some results for sieve estimation with dependent observations. In W. Barnett, J. Powell, & G. Tauchen (Eds.), Nonparametric and semi-parametric methods in econometrics and statistic. New York: Cambridge University Press. Williams, 1986, The logic of activation functions, Vol. 1, 423