Критерії аналізу логіко-лінгвістичних моделей речень природної мови

Abstract

Для здійснення змістовного аналізу електронних текстових документів запропоновано використовувати формальні логіко-лінгвістичні моделі. Метою статті є опис критеріїв аналізу формальних моделей, що здатні відображати зміст речень природної мови та формуються з використанням математичного апарату логіки предикатів. Описані критерії аналізу логіко-лінгвістичних моделей необхідні для побудови формальних моделей електронних текстових документів.The article describes the main text models used today as a tool for content processing electronic text documents. To make a content analysis author proposes to use formal logic and linguistic models, which are based on functional relationships between the principal and subordinate parts of natural language sentences. The article is to describe the criteria for analysis of formal models that can reflect the content of natural language sentences and which are formed using mathematical tools of predicate logic. For this purpose, the study researches principles of construction of logic and linguistic models of natural language sentences and formulates four criteria of analysis. First criterion analyzes the number of simple predicates in logic and linguistic model that helps to identify information about the type and composition of natural language sentences. The second criterion analyzes potency of set of predicate variables and constants of logic and linguistic model, which affects the number of simple predicates and identifies the type of individual forms of logic and linguistic model. The third criterion focuses on the analysis of logical operations that used in logic and linguistic model. That makes it possible to analyze the sequence of considerations referred to the natural language sentence. The forth criterion examines the presence of identical components in logic and linguistic models of natural language sentences from different sets of predicate variables and constants. Described analysis criteria of logic and linguistic models required to build formal models of electronic text documents using the mathematical apparatus of predicate logic.