Examination of Word-based Statistical Machine Translation Decoder

The first generation of the statistical machine translation was word-based, and its performance was low. More recently, phrase-based statistical machine translation has gotten better results.
In the proposed method, we thought the English-Japanese word dictionary and English-Japanese translation patterns were equivalent to the translation model of SMT. And we thought the word tri-gram was equivalent to the language model of SMT. Consequently, we thought the proposed method was equivalent to a word-based SMT decoder.