TRANSLATION-BASED CONSTRUCTION OF AN UZBEK WORDNET FROM PRINCETON WORDNET

Sobirov O. ., Xudayberganov B. . & Qodirberganova S. .


Annotatsiya

This study presents a translation-based methodology for constructing an Uzbek WordNet by leveraging the lexical and semantic structure of Princeton WordNet. As a low-resource language, Uzbek lacks comprehensive semantic networks required for advanced natural language processing applications, including word sense disambiguation, semantic search, and knowledge representation. The proposed approach operates at the synset level and involves translating lemmas, glosses, and semantic relations while preserving the original hierarchical structure. A hybrid pipeline combining automatic neural machine translation and manual refinement is employed to ensure semantic consistency. Special emphasis is placed on gloss-based disambiguation to address polysemy and improve translation accuracy. The study focuses on validating the feasibility of the approach rather than presenting large-scale evaluation results. Therefore, quantitative performance metrics are intentionally deferred to future research, where expanded datasets and expert-driven validation will be incorporated.


Adabiyotlar

  1. Bakay Ö., Ergelen Ö., Sarmış E., et al. Turkish WordNet KeNet. — Bogazici University, 2020.
  2. Agostini A., Usmanov T., Khamdamov U., et al. UZWORDNET: A Lexical-Semantic Database for the Uzbek Language. — Inha University, 2021.
  3. Madatov Kh. A., Khujamov D. J., Boltayev B. R. Creating of the Uzbek WordNet based on Turkish WordNet // AIP Conference Proceedings. — 2022. — Vol. 2432.
  4. Sobirov O., Allashkurov D., Rustamova D. Creating a Semantic Network of the Uzbek Language Based on English WordNet // Modern Science, Technology and Society. — 2026.
  5. Sobirov O., Allashkurov D., Rustamova D. Comparative Analysis of Expand and Merge Methods in Uzbek WordNet Construction // International Scientific Conference. — 2026.