To the Question of Developing an Information System for Processing Textual Information on the Basis of Semantic Groups

Abstract

The paper considers the information approach to the analysis of the text, based on splitting the sentence into its component parts and highlighting the topic and the rheme. The relevance of the work is justified by the need to search and identify useful information for the user, which he could use when performing the necessary scientific research. The concept of structural-semantic groups of sentences is introduced and requirements for them are defined. The sense group of a sentence is determined on the basis of an analysis of the relationships between the words of the sentence; and it includes some set of adjacent words that define some new image. The script of the analysis of the text information on the basis of the offered approach is described. A description of the approaches is provided for splitting the text of a document into semantic groups. The detailed results of the work of the program complex for various target installations for processing text are given. The presented results show the capabilities of the developed software package: the structuring of individual text sentences; formation of key words in the form of semantic groups for further analysis; selection of semantic groups that determine the main meaning of sentences and text; significant reduction of the text while preserving the semantic component. In the future it is supposed to expand the functionality of the complex and check the main ideas when processing large information arrays.