Approximation results in Orlicz space using max-min Fractional type neural network operators

In this study, we introduce and study the max-min fractional type neural network operators. The operators are generated by sigmoidal activation functions, which play an important role in the approximation process. These operators provide a constructive tool for the approximation of functions that may not necessarily be continuous. In particular, the main result proved here is a modular convergence theorem in Orlicz spaces. We provide some graphical representations to illustrate the approximation error of the function through activation functions.