Semantic Search Tools Based on Ontological Representations of Documentary Information

Autor: N. V. Maksimov, K. V. Monankov, A. A. Lebedev, N. A. Bal, S. G. Kyurcheva, O. L. Golitsina
Rok vydání: 2019
Předmět:
Zdroj: Automatic Documentation and Mathematical Linguistics. 53:167-178
ISSN: 1934-8371
0005-1055
DOI: 10.3103/s0005105519040046
Popis: Ontological means of identification and representation of text documents semantics in relation to the problems of interactive information retrieval are considered. The implementations of ontologies operations are presented, which allow forming images of new meanings in the subject area. Taxonomies of relations (paradigmatic and syntagmatic) and entities (polythematic thesaurus of concepts) are used to perform operations, as well as to identify fuzzy relationships (in the case when entities and relationships are specified at different levels of generality and/or expressed by different linguistic constructions). Interactive tools are proposed that use the operation of constructing aspect projections for graph representations of ontologies, which make it possible to reduce the dimension of the graph to a level acceptable from the point of view of display and perception. The possibilities of context-based use of entities and relationships for the development of the search process are considered. This paper discuses the use of ontology graphs of organizational processes as a “navigation map” that provides new entry points in the graphical interface, that is, objects that will be contextually defined patterns of search queries in similarity selection tasks or expert analysis tasks. This allows one not only to increase the completeness and accuracy of information retrieval, but also to make the scheme of search navigation more natural, bringing it closer to the schemes of understanding and synthesis of knowledge.
Databáze: OpenAIRE
Externí odkaz: