ONTOLOGY-DRIVEN MEANS FOR PROCESSING AND PRESENTATION OF LARGE ARRAYS OF UNSTRUCTURED TEXTS

Abstract

The subject of the article's research is methods of ontology-driven processing and presentation of unstructured texts in a global environment. The goal of the work is to improve the efficiency of automatic search, classification and selection of the necessary information contained in electronic publications in an unstructured form by developing a model, method and means of automated processing and transdisciplinary presentation of text documents created in Ukrainian, Russian and English languages. In accordance with the goal, the following tasks were set: to conduct an overview of models of processing and presentation of unstructured texts, to identify the advantages of information systems with an ontology-driven architecture, to develop a conceptual model and means for automated processing and transdisciplinary presentation of text documents, improve the method of recursive reduction, with the help of developed method and means to process a large array of unstructured texts (in particular information base of knowledge of scientific products (IBK SP), educational programs, as well as educational and research projects of the National Center "Junior Academy of Sciences of Ukraine"), to present them in a form that allows to reveal the quality and completeness of the knowledge system available in the databases and, thus, experimentally to confirm the effectiveness of the developed means. The conducted research is based on the following methods: systemological classification, algebraic-logical and axiomatic methods, the method of recursive reduction. The following results were obtained: the existing models of processing and presentation of unstructured texts were described, the technological aspects of the ontology-driven approach to the processing and presentation of large arrays of network documents were improved, the technology of forming ontologies of documents based on the representation of their taxonomies, connections between their contexts and multiple representations was developed, which provides effective search for information in unstructured texts. Conclusions: The advantages of ontology-driven means for processing and presentation unstructured texts are substantiated. The method of recursive reduction by constructing a transformation, which leads to the predicative form of ontologically given descriptions of information, gained further development. The idea of effective search, processing, classification and selection of the necessary information has gained further development.