String to Tree and Tree to String Statistical Machine Translation for Myanmar Language

Abstract

This paper contributes the first published evaluation of the quality of string-to-tree (S2T) and tree-to-string (T2S) statistical machine translation methods between Myanmar and Chinese, English, French, German in both directions. The performance of machine translation was automatically measured in terms of BLEU and RIBES scores for all experiments. In addition we performed a comparative study of the performance of phrase-based statistical machine translation (PBSMT) and T2S using human judgment. We found that the results obtained using the BLEU automatic evaluation metric were misleading and found that the T2S approach is suitable for distant languages to Myanmar machine translation.