Markov-based reordering model for English-Myanmar translation

In statistical machine translation, reordering is crucial component for translation of the different languages with different word orders. Without reordering during language translation, sentences can only be translated properly into a language with similar word order. An effective reordering scheme is essential to model translation between languages with different word orders, such as SVO-languages (English or Chinese) and SOV-languages (Japanese or Turkish or Myanmar). Our Language, Myanmar is a verb final language and reordering is needed when it is translated from other languages with different orders. In this paper, we focus on reordering rule generation and Markov-based reordering model implementation that can be incorporated into our English-Myanmar translation model. Parallel tagged aligned corpus is used as a resource for reordering rule generation and First Order Markov theory is applied for reordering model. Moreover, the alignment parameters for the reordering model are taken from the reordering rules automatically extracted from parallel tagged aligned corpus.