Fuzzy Analysis (Image Construction) of the Language Structure on a Finite Set of Insufficient Data
F. Criado, Tamaz Gachechiladze, Nato Jorjiashvili, Teimuraz Mandjaparashvili, Hamlet Meladze, Gia Sirbiladze, Tamaz Tsilossani, G. Z. Tsertsvadze · Journal of Quantitative Linguistics · 2004
Fuzzy logic and fuzzy theories were initially proposed to describe linguistic variables, i.e., to describe the meaning of words in natural language. Originally, Zadeh thought that the area of linguistic would be one of the major fields of application for this new formalism. Surprisingly, the main area of application is now control, and in comparison with control there are only a few application in the field of linguistic. In view of it, this work describes a new approach to the study of natural language. The paper is devoted to numerical modeling of fuzzy organization of various language structures by using fuzzy distributions which contains data about important regularities controlling informational and psycho-physiological processes which accompany the generation of verbal language or printed texts. These distributions are characterized by so called linguistic spectrum describing various distribution mechanisms of marked out language elements over their sequences. Models of such linear structures are based on the assumption about the superposition of two kind of uncertainties: probabilistic and possibilistic.