Voronovskaja type theorems and high-order convergence neural network operators with sigmoidal functions (Q2178844)

The authors provide an asymptotic formula for neural network (NN for short) operators which are given in terms of sigmoidal functions, i.e., real functions satisfying meaningful assumptions (Theorem 3.1). Also, the authors describe an asymptotic behavior of a finite linear combination of NN type operators (Theorem 4.1). The article is endowed by a few applications with explicit sigmoidal functions as the logistic and the hyperbolic tangent functions. The sixth section focuses on numerical results.