Work

A Synset-Centered Data Model for Uzbek WordNet Construction

Year: 2026

Type: conference-paper

Abstract: This work describes the process of converting the UzWordNet-GoldCore dataset from a table format into a structural model centered on synsets. GoldCore consists of 11,178 entries. During the conversion process, lexical units, senses, synsets, examples, semantic relations, and diagnostic states were created as separate objects. As a result, 7,072 lexical entries, 10,533 senses, 2,740 candidate synsets, 11,095 unique examples, and 9,086 relation objects were generated. Additionally, 213 domain conflicts and 125 ontology-category inconsistencies were assigned for expert review. Of the 9,086 relation targets, 2,500 were precisely mapped to a single synset, accounting for 27.51%. Among hypernym targets, only 65 out of 2,740 were precisely resolved. Unresolved and ambiguous targets were not automatically modified but were retained for expert review. This work evaluates not the results of downstream tasks but the resource's structural design, monitoring, and quality control processes.

Source: International Conference on Computer Science and Engineering (UBMK)

DOI: 10.1109/ubmk71847.2026.11745189

Authors: Botir Elov Boltayevich, Kurbanova Mukhabbat Matyakubovna, Zilola Yuldashevna Xusainova, Aynura Akhmedova, Sitorabonu Tursunboyeva Xusniddin Qizi, Sarvinoz Mardonova Oripovna, Gulasal Khayrulloeva Fayzullo Qizi, Mutalibova Saodat O'Lmas Qizi

Institutions: National University of Uzbekistan, Bukhara State University, Buxoro davlat pedagogika instituti

Language: English

Topic: Natural Language Processing Techniques

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W7224870909

FWCI:

Cites: 12

Cited by:

Related to: 10