Neural dependency parsing of uzbek texts based on the universal dependency treebank

Matlatipov, Sanatbek, Fayzullayev, Xurshid

Al-Farg'oniy avlodlari · 2025-yil

Annotatsiya

The Uzbek language is morphologically rich with numerous affixes, and meaning in sentences is expressed through various suffixes of nouns, verbs, and other parts of speech.  In this article, we present a neural semantic analysis of Uzbek language texts based on a set of trees created using the Universal Dependencies (UD) corpus. This corpus contains 686 sentences and 7,950 tokens, with an average of 11.6 tokens per sentence. We construct a graph-based parser that employs a two-stage architecture - a BiLSTM-based contextual encoder combined with a biaffine function that integrates head and dependent projections.

Maqola ma’lumotlari
MualliflarMatlatipov, Sanatbek, Fayzullayev, Xurshid
JurnalAl-Farg'oniy avlodlari
Nashr sanasi2025-10-08
Son3
Betlar135-138
TilRus

Kalit so‘zlar

корпус универсального дерева зависимостей, Universal Dependency treebank, Uzbek language, kam resursli tillar universal bog‘liqliklar, o‘zbek tili, neyron semantik tahlil

Ilmiy soha

Al-Farg'oniy avlodlari jurnalidan boshqa maqolalar

Al-Farg'oniy avlodlari — barcha maqolalar