We developed English-Bangla parallel corpora for statistical machine translation. By hand we tagged 20,000 words of our Bangla corpus according to their particular part of speeches. In our work we also suggested a method ...