One Approach to the Description of Linguistic Uncertainties

Abstract

The basic concept of this paper is to interpret some uncertainties of our speech. It differs from traditional approach [1]. The main idea is to build new functions using the Fuzzy C-means Clustering Algorithm [4]. We will adhere to Zadeh’s idea to represent a hedge as an operator acting on fuzzy sets and will consider four uncertainties: much larger, much smaller, slightly smaller and slightly larger (we will call them modifiers). These modifiers act on two fuzzy sets, so we have a linguistic variable with two values. The result of research is an algorithm, which determines the membership function of this variable’s value after the modifier’s action. Some properties, showing the conformity of the algorithm to the semantics of natural language and convergence in particular cases, are also described. The application of this work is the task of extracting from the database objects satisfying the semantics of the modifiers used.