Principles For Creating An Ontological Base For Mathematical Terms In English–Uzbek Machine Translation

Abstract

Mathematical terminology often presents significant semantic ambiguity in machine translation (MT) systems, especially in language pairs with structural and conceptual asymmetry such as English and Uzbek. This paper discusses the linguistic and technological principles for creating an ontology-based lexical base aimed at enhancing the translation accuracy of mathematical terms. The proposed ontological model integrates linguistic analysis, conceptual structuring, and semantic relations to ensure consistent and contextually appropriate translation of polysemous terms. A bilingual corpus of English–Uzbek mathematical texts was used for term extraction, normalization, and semantic mapping within an ontology framework developed in Protégé using OWL/RDF representation. The research findings demonstrate that the ontology-based approach significantly improves semantic coherence and reduces contextual translation errors compared to conventional neural MT models. The proposed principles contribute to the development of a domain-specific linguistic base for English–Uzbek machine translation and can serve as a foundation for future multilingual ontological systems in scientific and technical fields.