What are the best ways to identify clauses in a sentence? I would like to not only extract clauses from a sentence but also identify the type of clause it is. i.e. dependent or independent clause.
A clause is grammatically defined as the part of the sentence having a subject and predicate. For example consider the sentence:
"In fact, the last time he tried to lose weight, he ended up actually gaining weight."
"He tried to lose weight" is a dependent clause while,
"He ended up actually gaining weight" is an independent clause.
I have come across some results from CONLL but these are pretty old (2001), so I was wondering whether anyone has an idea to some better work done on this subject.
If you have annotated corpora, you can use a learning approach.
Or you can implement rules defined by linguistic experts.
But in general, you have two steps to do: 1) clause boundary identification and 2) clause type identification (here you can use binary classification)
@Iskandar, thanks for listing the steps to achieving this task. I am looking for exactly these linguistic rules which you have mentioned. Can you cite some work or links where I can find some information on this? Thanks.
@Osman, as the Stanford group is pretty big, can you please give me a link to the person/team at Stanford that is working on this or better, some paper(s) or work that is related to this? Thanks.
Iskandar Keskes, Farah Benamara and Lamia Hadrich Belguith, «Clause-based Discourse Segmentation of Arabic Texts», The eighth international conference on Language Resources and Evaluation (LREC 2012), Istanbul, 21-27 may 2012.
Iskandar Keskes, Farah Benamara and Lamia Hadrich Belguith, «Discourse Segmentation of Arabic Texts Based on Cascade Grammars», NooJ conference (NooJ 2012), Paris, 14-16 june 2012.
Http://nlp.stanford.edu/
http://nlp.stanford.edu/projects/arabic.shtml
natural language processing group
Thanks for all the suggestions everyone! I'm going to take a couple of days to go through all the links and material you have suggested. I'm sure I'll be able to find a nice way to solve the problem with this help. Thanks!
- Badges
- Science topic