A fuzzy knowledge-based approach to the alternative classification problems

Presents a fuzzy knowledge-based approach to the problem of classifying alternatives in incomplete and imprecise real-world environments. Expert domain knowledge is modelled by a fuzzy conjunctive production rule base. By applying a comparison ratio function, defined in this paper, we propose a new reasoning algorithm to identify the most suitable rule from the pre-defined production rule base and to infer the class of an input observation. Unlike traditional similarity (or matching) functions, this comparison ratio function can be associated with a reasoning threshold value to prevent mis-firing of the consequent portion of a classification rule. The fuzzy knowledge-based approach proposed in this paper is an efficient approach to classify classes of alternatives