With our system, we aim to reduce the number of unknown words and ungrammatical sentences. Thus, we analyze the outputs according to these factors.