Word alignment system based on hybrid approach for Myanmar-English machine translation

Word alignment is a basic and critical process in the Statistical Machine Translation (SMT). Word alignment is to identify word correspondence that are translations of each other based on information found on parallel text. Essential for building parallel corpora is the alignment of translated segments with source segments. A parallel corpus is a collection of texts in two languages, one of which is the translation equivalent of the other. Nowadays, Myanmar-English word-aligned parallel corpora are not available. This paper describes the construction of an aligned Myanmar - English parallel corpus to be able to use as a resource in Myanmar-English machine translation. The proposed system uses the combination of corpus-based approach and the dictionary lookup approach. The corpus-based approach is based on the first three IBM models and Expectation Maximization (EM) algorithm. For the dictionary lookup approach, the proposed system uses the bilingual Myanmar-English Dictionary. The system also uses a list of cognates and morphological analysis to get better alignment accuracy. Accuracy of modern statistical machine translation depends on good word alignment.