Exploring the Use of Target-Language Information to Train the Part-of-Speech Tagger of Machine Translation Systems