Abstract
With the rapid growth of disinformation, cognitive manipulation and coordinated information campaigns in digital media, there is a need to develop intelligent methods for identifying and analyzing information operations. This paper proposes a conceptual model that integrates a multilingual annotated corpus, an ontological knowledge base and a semantic knowledge graph for the systematic study of mechanisms of information impact.The methodology of the research includes the construction of a specialized framework formed out of messages collected from Telegram channels, news portals and social networks. The data goes through a multi-level annotation using the Label Studio platform, where experts manually mark key entities, including military terms, target audiences, sources, actors, and emotional evaluations. The annotated corpus is semantically corresponded with the ontology of the subject field, formalized in OWL and enriched with the military thesaurus MIL_term, which provides consistency of terminology and support for multilingual analytics.The ontological model is transformed into an RDF-graph of knowledge, reflecting the relationships between entities, events, tactics and narratives. SWRL-rules are used to identify hidden patterns, and the developed SPARQL-queries allow to extract complex analytical patterns, including chains of "actor – tactic – narrative – audience". The proposed approach forms the basis for complex analysis of information flows, early detection of threats and construction of analytical scenarios, which makes it applicable for research and monitoring of information operations in multilingual digital environments.