COMPARATIVE ANALYSIS OF MACHINE TRANSLATION SYSTEMS USING THE METEOR METRIC

This study evaluates the accuracy of several machine translation systems using the METEOR metric. The proposed Tilmoch system is compared with Google Translate, Yandex, Microsoft, and DeepL using source, reference, and candidate translations. The results show good translation quality across the evaluated systems, while UzMarine achieved a higher score due to its hybrid approach, which combines rule-based linguistic processing with a Transformer-based model. Experiments were conducted on a parallel corpus containing 142,838 Uzbek-English sentence pairs and vice versa. The findings demonstrate the effectiveness of the proposed hybrid approach for improving Uzbek–English machine translation accuracy.

Hali tarjima qilinmagan