The Use of Model-Theoretic Methods for Extracting Ontological Knowledge in the Domain of Information Security

D. E. Palchunov , e-mail: palch@math.nsc.ru, P. A. Stepanov , e-mail: stefan.nsk@gmail.com

The paper is devoted to the methods for extracting on-tological information from natural language texts. We give a model-theoretic formalization of relations between concepts such as the synonymy relation, the hyponym-hypernym relation, the relation "one concept is used to define another concept". Methods for extracting these relationships, as well as extracting definitions of concepts from natural language texts, are presented. A language for describing linguistic patterns is used for natural text analysis. The developed software system showed very high accuracy and completeness of the extracted knowledge. The obtained results are used in the development of an expert system for information security and risk management systems for information security.

Keywords: subject domain ontology, information security, model-theoretic methods, the logical analysis of natural language, automata theory, non-deterministic automata, natural language processing, knowledge extraction, extraction of definitions of concepts

pp. 8–16