Talk · International Conference on Computer Science and Engineering · September 1, 2026

Development of an Uzbek Paraphrase Tool for Analyzing Literary Aphorisms and Cultural Terms Based on Uzbek Travelogue and Film Discourse

Zarnigor M. Khayatova, S T Ismailova, Kumushkhon Razokova, Rakhimova Gulchekhra, Nigina Barotova, Aziza Naimova

Abstract

This paper presents a web-based Uzbek paraphrase tool for identifying and interpreting paraphrase transformations in literary aphorisms and culturally marked expressions from Uzbek travelogue and film discourse. The corpus contains 14,000 Uzbek sentences organized into 7,000 aligned source–paraphrase pairs. The study focuses on a subcorpus of 3,700 pairs, including 1,600 travelogue pairs and 2,100 film-discourse pairs. The system applies text normalization, tokenization, corpus-based candidate retrieval, term frequency–inverse document frequency cosine similarity, changed-unit alignment, part-of-speech annotation, transformation classification, and color-based visualization. Paraphrasing is modeled through semantic invariance, linguistic transformation, context, and style. The main processing challenges involved agglutinative suffix variation, flexible word order, multiword cultural expressions, and figurative aphorisms. The results indicate that corpus-based and linguistically interpretable processing can support explainable Uzbek paraphrase analysis.

Natural Language Processing TechniquesTopic ModelingLinguistic research and analysisparaphraseUzbek languagecultural terms

Development of an Uzbek Paraphrase Tool for Analyzing Literary Aphorisms and Cultural Terms Based on Uzbek Travelogue and Film Discourse

Abstract

More by Zarnigor Khayatova