Semantics based English-Arabic machine translation evaluation

Some classic machine translation (MT) Evaluation methods, such as the bilingual evaluation understudy score (BLEU), have notably underperformed in evaluating translations for morphologically rich languages like Arabic. However, recent remarkable advancements domain of word vectors and sentence opened up new research avenues low-resource languages. This paper proposes a novel linguistic-based method English-translated sentences The proposed approach includes penalties based on length, positions, context-based schemes part-of-speech tagging (POS) multilingual sentenceBERT (SBERT) models evaluation. technique is tested using pearson correlation performance parameter compared with state-of-the-art techniques. experimental results demonstrate that model evidently outperforms other MT methods BLEU.