Approximation Properties of the Kantorovich Type Exponential Sampling Neural Network Operators

In this article, we define a Kantorovich version of exponential sampling neural network operators, which employ sigmoidal functions as activation functions and examine its approximation capabilities. We establish theorems concerning convergence and estimate the rate of approximation for the class of logarithmic continuous functions and ‐integrable functions in Mellin sense. Moreover, we determine the convergence rate in the modulus of continuity. Afterward, we give several examples of activation functions and provide the numerical verification of our results.