Title :
The best approximation properties and error bounds of Gaussian networks
Author :
Liu, Binfan ; Si, Jennie
Author_Institution :
Dept. of Electr. Eng., Arizona State Univ., Tempe, AZ, USA
Abstract :
The best approximation of any C2 function with support on the unit hypercube Im in Rm is considered in the present paper. We prove that a Gaussian radial basis network with centers defined on a regular mesh in Rm has the best approximation property. Moreover, an upper bound (O(N-2)) of the approximation is obtained for a network having Nm units
Keywords :
computational complexity; error analysis; feedforward neural nets; function approximation; Gaussian radial basis network; best approximation; error bounds; unit hypercube; upper bound; Feedforward neural networks; Gaussian approximation; Green´s function methods; Hypercubes; Integral equations; Measurement standards; Neural networks; Nonhomogeneous media; Radial basis function networks; Vectors;
Conference_Titel :
Decision and Control, 1993., Proceedings of the 32nd IEEE Conference on
Conference_Location :
San Antonio, TX
Print_ISBN :
0-7803-1298-8
DOI :
10.1109/CDC.1993.325705
