Ways of Recovery of Text Information Represented in the Form of a Logic and Linguistic Model

Abstract

The materials of the article substantiate the urgency of solving the problem of identifying meaningful links in electronic text documents in order to further compare their content and improve the operation of plagiarism detection systems. An important step is to assess the reliability of the formed formal models. Therefore, the aim of this article is to study the algorithm of automatic analysis of logic and linguistic models of electronic text documents for the reproduction of textual information, which combines the basic properties of the text and its components. The logic and linguistic model of a text document reflects the main relationships between structural components; it is an ordered quadruple and an array of logic and linguistic models of sentences of natural language, which are included in the text. The author proposes several ways of restoration of textual information, starting from the structure of the logic and linguistic model of an electronic text document, which contains a linguistic and semantic-syntactic component. The article describes the schemes of text information recovery, it chooses the combined method, which provides the analysis of semantic-syntactic component in parallel with the analysis of the text base, in particular, its components - a set of sentences containing connections between logic and linguistic models of sentences of the text within electronic text document. It has been developed an algorithm for recovery of textual information presented in the form of a formal logic and linguistic model of an electronic text document, and there have been described the stages of abovementioned algorithm. All steps of the algorithm are demonstrated on the example of analysis of a specific given logic and linguistic model of a fragment of an electronic text document. The author conducted experiments on the restoration of textual information for scientific style texts. During experiments, it was revealed, that the main factors influencing the restoration of textual information include the removal of homonymy, as well as different interpretations of synonymous constructions and invariant forms of logic and linguistic models of sentences of natural language.