Velychko V. Scientific and technological foundations of knowledge-oriented processing of natural language texts and its application

The dissertation is devoted to solving the scientific and applied problem of increasing the efficiency of ontological analysis, consolidation and extraction of knowledge from natural language texts in conditions of processing a large amount of information from heterogeneous subject areas. Conceptual and methodological bases of processing and analysis of natural language texts and construction of ontologies of subject area are developed, which represent the basis for development of analytical and practical methods of formation, specification and systematization of knowledge on subject area on the basis of semantic analysis and methods of knowledge extraction from natural language texts. Distinctive features of the concept of natural language processing are the ability to synchronize the sources of texts in natural language with the contexts of concepts from the repository of ontologies, which reflects the current state of the subject area, and the use of growing pyramidal networks to reflect the entities and relationships in ontologies. Growing pyramidal networks allow the use of high-speed graph databases to solve problems of information interaction. The class-oriented method of building a growing pyramidal network using parallel computing and agent technologies has been improved. The method is based on optimized algorithms of its structural reconfiguration and allows to build a growing pyramidal network of invariant structure before ordering the source data. A local statistical method of concept extraction in growing pyramidal networks has been developed, which makes it possible to form concepts in the notation of the propositional logic of minimum length with maximum support in polynomial time. A network platform for the implementation of logical-linguistic analysis of information resources and the formation of network-centric interactive knowledge systems based on its results has been created. Based on the use of ontologically controlled tools of the created basic cognitive IT technology, a number of network-centric interactive systems of knowledge in the areas of education, museum affairs, health care, information-analytical and practical activities of institutions in various fields have been implemented.