Approximation by a class of neural networks to integrable function
Feilong Cao · 2007
The estimates of approximation rate for Cardaliguet-Eurrarad type neural networks operators approximating the continuous functions over the real line were given by the modulus of continuity.At the same time,in the case of approximating Lebesgue integrable function,we construct the correspondent neural network operators,and use Jackson type inequality to determine its rate of approximation.