An approach to Lao-English rule based machine translation

This paper presents an approach to Lao-English rule based machine translation. In our approach, we start by modifying the Lao word segmented algorithm to mark the name entities part, which will be reordered according to its type later. Then, using the transfer-based strategy, the sentence is analyzed to build the source dependency structure, which is then transferred to the target dependency structure. The target text is then generated according to its attribute information. The machine translation performance obtained from the proposed method, in BLEU, METEOR, ROUGE-L, was 0.5920, 0.5482 and 0.8586, respectively. These results are better than those achieved by Google Translate.