Engineering Physics Annotation << Back A MACHINE LEARNING APPROACH FOR FUZZY CLASSIFICATION MODELS D.V. POLYAKOV,A.I. YELISEYEV, D.S. ANDREYEV, A.YU. SELIVANOV The paper formalizes the problem of fuzzy classification in a general form. A fuzzy mathematical model is proposed that combines expert assessments and a training sample. Based on proposed model, a variational problem is set to find the form of membership functions and fuzzy logic operations that will minimize the model error. The variational problem is reduced to an optimization problem by using parametric norms and conorms and a Fourier series to represent an arbitrary function with specified constraints. A method for solving the optimization problem is proposed and justified, which is an analog of machine learning of an artificial neural network by the method of backward error propagation. The choice of the initial parameter assignment proposed and justified, and the initial view of the corresponding membership functions is constructed and presented. Keywords: classifi cation, fuzzy logic, fuzzy set theory, machine learning, variational problem, optimization problem. DOI: 10.25791/infizik.7.2021.1217 Pp. 19-35.