Improved method of increasing the accuracy of selection of relevant text messages based on the identification of referential coherence of the text elements

The selection of text messages based on keywords is ineffective because it ignores the context of the use of the concept and its semantic relationships. The main perspective concept of processing NL-texts is a semantic network. World Wide Web Consortium approved and advances the project of the Semantic web on the basis of the concept of a semantic network. The main reason that hinders the active development of the Semantic web is the lack of implemented methods for semantic markup of existing text documents. In this regard, the article highlights the conceptual approach consistent conversion of texts to the level of the discourse graph, it will lead to a semantic formalization of the text and automatic processing of NL-texts. Within this approach we propose an enhanced method of selecting relevant text messages based on the identification of referential coherence of the text elements.