Learning word alignment from nltk

Question

I have a parallel corpus for english-german. Is there a way to extract word alignment table from this corpus using nltk? I don't know if nltk.align is supposed to do this. I am unable to figure out from the documentation.

Answer 1

Look at the source of the modules in the nltk.translate package (previously known as nltk.align); you'll find descriptions of the available algorithms and references to the research literature that explains them in more detail.