General wavelet transform and application in neural networks. (Chinese. English summary) Zbl 1037.94559
Summary: A new kind of wavelet transform relating to the problem of the realizability of \(f\in L^p (l^n)\) by single hidden layer feed forward artificial networks is introduced and its properties are studied. As a corollary, another constructive proof of Cybenko’s theorem is presented.
MSC:
94C05 | Analytic circuit theory |
42C40 | Nontrigonometric harmonic analysis involving wavelets and other special systems |
92B20 | Neural networks for/in biological studies, artificial life and related topics |