English–Uzbek Machine Translation based on Independent Parts of Speech and New Mathematical Models
Abstract
This article is devoted to the development of new mathematical models based on independent parts of speech in the process of translating from Uzbek into English. The study analyzes the construction of words belonging to Uzbek independent parts of speech through affixes and examines new mathematical models built on an expandable input language, implemented through software tools. The research provides a detailed analysis of the agglutinative morphology of the Uzbek language, word formation processes, and affixation system, and compares them with the analytic features of English. The proposed model integrates morphological, syntactic, and semantic stages and introduces new mathematical approaches for processing both simple and complex words. The Machine translation system developed on the basis of these models effectively addresses grammatical differences between the two languages and ensures high-quality and semantically accurate results. In addition, the paper presents the software code and visual interface of the system, developed using 67 mathematical models, as well as the architecture, components, and implementation stages of the automatic translation system. This approach not only enables an in-depth analysis of the complex morphological structure of the Uzbek language but also ensures precise and systematic outputs in translation into English. The results of the study demonstrate practical applicability in the field of scientific translation and hold value for young researchers, higher education institutions, and international organizations.