LINGUISTIC MODIFIERS AND IMPRECISE CATEGORIES

We study knowledge-based systems using fuzzy logic and we focus on the representation of knowledge through linguistic variables characterized by means of fuzzy qualifications or labels. We study a new form of linguistic modifier which slightly changes the qualifications through either a weakening or a reinforcement. This modifier is an important tool for approximate reasoning for two reasons: its use is equivalent to a simple rule given in a symbolic way, avoiding computations and compatible with the fuzzy logic; it enables gradual rules to be used in the context of deduction rules and corresponds to the idea of gradual changing of category.