Approximation results on neural network operators of convolution type

Abstract

In the present paper, we introduce three neural network operators of convolution type activated by symmetrized, deformed and parametrized B-generalized logistic function. We deal with the approximation properties of these operators to the identity by using modulus of continuity. Furthermore, we show that our operators preserve global smoothness and consider the iterated versions of them. Here, we find it is worthy to mention that these operators play important roles in neural network approximation since most of the basic network models are activated by logistic functions.