Greetings, Thanks a lot for this paper, I will look it over and then ask any remaining questions. I am, indeed, at the University of Washington. I was hoping to discover that you had used the dependency structures to extract semantic information to build your ontology. I will read the paper in hopes of uncovering more on that.
Thank you so much, Caleb On Mar 2, 2010, at 1:13 AM, Kais Dukes wrote: > Hello Caleb, > > Thank you for your interest in the Quranic Arabic Corpus. You've also > raised lots of useful questions. I've attached a recent PDF paper, > which has been accepted as part of the INFOS 2010 conference. > > =========================== > > -- A Dependency Treebank of the Quran using Traditional Arabic Grammar -- > Kais Dukes (University of Leeds) and Tim Buckwalter (University of Maryland) > > Abstract: > The Quran is a significant religious text, followed by the 1.5 billion > believers of the Islamic faith worldwide. The text dates to 610-632 CE > and is written in Quranic Arabic, the direct ancestor language of > modern standard Arabic in use today. This paper presents the Quranic > Arabic Dependency Treebank (QADT) and reports on the approaches and > solutions used to apply Natural Language Processing to the unique and > challenging language of the Quran. This project differs from other > Arabic treebanks by providing a deep computational linguistic model > based on historical traditional Arabic grammar (إعراب القرآن الكريم). > The treebank is part of the Quranic Arabic Corpus > (http://corpus.quran.com), a popular free Arabic resource developed at > the University of Leeds. Motivated by the importance of the Quran as a > central religious text, we also report on how online collaborative > annotation was used to bring together Quranic scholars and Arabic > language experts to ensure a high level of accuracy for grammatical > analysis of the entire Quran. > > =========================== > > I assume your that department is at the University of Washington? I > hope that the attached paper answers most of your questions about the > project. If not, please feel free to e-mail the mailing list > (comp-quran@comp.leeds.ac.uk). I would be happy to further clarify > anything beyond what is detailed in the above paper. There are also > two LREC papers which are forthcoming (I can send these over in a > couple of weeks) that also detail different parts of the project. > > Kind Regards, > > - Kais Dukes > > Language Research Group > School of Computing > University of Leeds > > http://corpus.quran.com - The Quranic Arabic Corpus > comp-quran@comp.leeds.ac.uk - Computational Quranic Arabic discussion list > > >> On Tue, Mar 2, 2010 at 5:48 AM, Caleb Barr <newcreat...@gmail.com> wrote: >> Greetings, >> >> I had a couple quick questions on the Quranic Arabic Corpus: >> >> Would you please refer us to the paper detailing the process by which you >> constructed the treebanks? I saw that you mentioned a Nivre (2009) paper on >> slide 27 on your phd-presentation.ppt. That slide mentions two approaches, >> machine learning and hand-written rules. How did the performance of each >> approach compare? Have you used dependency grammar to aid in the >> construction of your ontology? What role does dependency grammar (and which >> specific implementations / algorithms) play in the scope of the whole >> project? I'm going to talk to some people in my department about this >> project on Friday. >> >> Thanks, >> >> Caleb Barr > <qadt-infos2010.pdf>
