Tectogrammatical trees and fuzzy natural logic in linguistic characterization of dynamic processes

摘要

This paper is a contribution to the applications of fuzzy logic in modeling of linguistic semantics. The topic of interest is description of the past course of a dynamical process that is given by data in the form of time series. We suggest to use the well developed technique of linguistic theory, the so called tectogrammatical trees using which linguists represent semantics of sentences. The core of the suggested procedure is construction of the tectogrammatical tree that is then assigned a formula of fuzzy natural logic. As the latter construes semantics of the given sentence, we can construct a model on the basis of the given data. If successful, the initial sentence can be taken as a linguistic characterization of them. In this paper we demonstrate how sentences characterizing the course of the trend of time series can be assigned the formulas of fuzzy natural logic expressing their semantics via tectogrammatical trees.