Talk · International Conference on Computer Science and Engineering · September 1, 2026
Building a Phraseological Ontology of Uzbek Idioms and Evaluating Similarity Measures for Their Detection in Text
Nazirova Elmira Shodmonovna, Nilufar Z. Abdurakhmonova, N. Madetbayeva
Abstract
This paper presents a phraseological ontology for Uzbek idioms. The ontology classifies idioms along five dimensions: semantic, structural, stylistic, thematic, and translation equivalence. The lexical layer of the ontology, intended for idiom detection, contains 2,098 idioms and 1,104 lemmas. The ontology was built in the Protege environment and stored in the Web Ontology Language format. To assess the practical value of the ontology, the task of detecting idioms in text was considered. Five word-overlap similarity measures were tested on a manually annotated, stratified gold set of 400 sentences drawn from eight works of Uzbek literature. According to the results, the best measure reached F1 = 0.21 on the full 400-sentence set; even when the evaluation was restricted to the cases where the idiom is present in the ontology, the result did not exceed F1 = 0.37. Four reasons for this were identified, and it is substantiated that word overlap is a necessary but insufficient stage of idiom detection.
More by Nilufar Abdurakhmonova
- A model of a universal linguistic knowledge graph for Turkic languagesArctic XXI century · 2026
- Machine Learning-Based Topic Classification of Economic and Financial Texts in UzbekInternational Conference on Computer Science and Engineering · 2026
- Developing Morphological and Syntactic Analysis Models for Karakalpak as a Low-Resource Turkic LanguageInternational Conference on Computer Science and Engineering · 2026
- WordNet as a Model for Developing a Bilingual Thesaurus of English and Uzbek Qualitative AdjectivesInternational Conference on Computer Science and Engineering · 2026