[languagehacker]

Cool idea, but basically reinvents chunking, which NLTK already has (http://nltk.org/api/nltk.chunk.html). Keep up the writing and NLP research though :)