Lao-Thai machine translation using statistical model

摘要

This paper aims to present the Lao-Thai machine translation (MT). In order to develop the system based on statistical model, the language model was constructed. Due to Lao is a non-segmented language, the input text in Lao language was segmented into sequence of words using syllable longest matching. These segmented words were used to look for Thai corresponding words from the Lao-Thai dictionary and generate all possible translation sentences. Finally, the system selected the most probability of translation using the language model. The translation accuracy of the proposed system was 88.64%. The result shown that, this system outperforms the translation by Google Translate.