Моделювання залежностей за допомогою нечіткої бази знань з нечіткими регресійними рівняннями

Abstract

New structure of fuzzy knowledge base is suggested, in which the antecedent are presented in the form of fuzzy set terms, and the consequent in the form of linear relations "Input - output" with fuzzy coefficients. An algorithm of the inference of the knowledge base for the case of clear and fuzzy input data is proposed. Based on the theory of fuzzy identification mathematical statement of learning objectives of the proposed knowledge base on experimental data is formulated