---------- Forwarded message ----------
From: Kais Dukes <dukes dot kais at googlemail dot com>
Date: Fri, Dec 25, 2009 at 5:52 PM
Subject: Feedback required - Ontology of Quranic Concepts - and Pronoun Resolution + Named Entity Recognition
To: Kais Dukes <dukes dot kais at gmail dot com
Apologies for the mass anonymous email, but we are still in the
process of collecting names for the mailing list. I would really
appreciate people's feedback on three new features I am planning to
add to the corpus.
(1) Pronoun resolution. So this will indicate what a pronoun (e.g. he,
she, it, them) refers to.
(2) Named entity recognition. Mapping individual words (or spans of
words) to the concepts that they refer to.
(3) The ontology of concepts. Knowledge representation - lists all the
concepts in the quran, and shows how they are logically related (e.g.
ishmael - son-of - ibrahim).
I've not uploaded the new version of this website yet. I wanted to get
people's opinions. The ontology I have constructed so far has 300
concepts and about 400 relations. I would estimate a full ontology for
the quran would have several thousand concepts.
If you have the time, I would very much appreciate it if you could
look at these 8 screenshots that better explain the idea:
I'm really looking to feedback on this new proposed feature to the
Quranic Arabic Corpus. In terms of timing, I plan to mark up pronouns
and other named entities during the process of sytanctic annotation,
so as we build the treebank. I don't think that this will introduce a
large overhead to the project (in fact, its probably quicker if we do
this as we go along with syntax as opposed to during another seperate
The ontology outlined in the above screenshots focuses on named
entities in the quran (people, places, animals, etc). We can later
extend this to include an ontology section for legal topics (Islam
law) and also for faith-related issues.
Please let me know what you think of the above link.
-- Kais Dukes
School of Computing
University of Leeds