A HYBRID RULE-BASED AND TRANSFORMER APPROACH FOR UZBEK-ENGLISH MACHINE TRANSLATION: EVALUATION USING METEOR
Abstract
This article presents how hybrid translation tool works. Specifically, it indicates rule-based translation's IDF0 and also gives information about our tranlation system dataset which is done based transformer model. Moreover, it adduces the output evaluation of machine translation quality using the METEOR metric, with comparing the output of the English-Uzbek translation tool against human-produced translations. The study discusses how evaluation methodologies for machine translation systems work in bilingual contexts, emphasizing the importance across diverse domains. With the rapid enhancement of technology, there are different translation tools, each improving translation precision and fluency. This research provides a detailed analysis of translation evaluation metrics, specifically METEOR, applied to sentence-level translation tasks from English to Uzbek. It shows how low-resourced language is translated in our translation and how it is close to human translation.