ณัฐพงศ์ เนตรจินดา. Thai-English Statistical Machine Translation. Master's Degree(Computer Engineering). King Mongkut's University of Technology Thonburi. KMUTT Library. : King Mongkut's University of Technology Thonburi, 2010.
Abstract:
This thesis proposes Thai-English Machine Translation using statistical method. The
proposed system consists of two major processes, training process and translating
process. Training process is the process which prepares resources used in translation
process from bilingual corpus. The resources to be prepared are Translation Model,
which is the dictionary of Thai-English phrase pairs and probabilities of those phrase
pairs, and Language Model which is the collection of sequences of English words and
their co-occurrence probabilities. Translation process uses Translation Model to find the
translation of Thai phrases from input considering the phrase pair's probability. English
phrases are rearranged to create the proper sentence using word sequences from
Language Model considering co-occurrence probabilities. The developed system is then
evaluated using BLED score. The corpus used in the experiment is of the size 3,604
sentence pairs. This corpus is divided into 2 parts, 3,504 sentence pairs for training and
100 sentence pairs for testing. The results from the experiment show that Thai-English
Statistical Machine Translation yields better results compared to Google Translate and
Bing Translator.