[Corpora-List] A dependency parser for Arabic

[Yuval Marton]

Jack,

You might want to check out the Columbia CATiB parser (same group who developed Amira)

http://www1.ccls.columbia.edu/~ymarton/#_Teaching 
(resources and tools section)

It is one of the best dep parsers for Arabic to date, just evaluated in the EMNLP 2013 SPMRL shared task. 

I can provide you with more details if you email me directly. 

-Yuval 

--- Pardon typos, sent from my phone --- 

On Sep 9, 2013, at 4:23 PM, Kevin Gimpel <kgimpel at cs.cmu.edu> wrote:

> Hi Jack,
> TurboParser (http://www.ark.cs.cmu.edu/TurboParser/) includes a pretrained model for Arabic. (Not sure how the AMIRA tokenization differs from the tokenization of the CoNLL-X data used to train this model, but others might know.)
> The Stanford parser (http://nlp.stanford.edu/software/lex-parser.shtml) also has an Arabic model. You can get dependencies from the phrase structure parses, though not typed dependencies (http://nlp.stanford.edu/software/parser-arabic-faq.shtml#j).
> Kevin
> 
> 
> On Mon, Sep 9, 2013 at 4:12 PM, Jack Alan <j.o.alan2012 at gmail.com> wrote:
>> Hi eveyone,
>> 
>> I wonder if someone came a cross a dependency parser for Arabic. I've no access to any resources provided by LDC, so I'm looking for something **opensource**, i.e. free.
>> 
>> By the way, I'm using AMIRA[1] to perform tokenization. So, I want to feed the tokenized text into the dependency parser to do the job.
>> 
>> Could anyone pinpoint me to the proper tool to use, if any?
>> 
>> Jack
>> 
>> 
>> Ref:
>> [1] Diab, Mona. "Second generation AMIRA tools for Arabic processing: Fast and robust tokenization, POS tagging, and base phrase chunking." 2nd International Conference on Arabic Language Resources and Tools. 2009.
>> 
>> 
>> _______________________________________________
>> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
>> Corpora mailing list
>> Corpora at uib.no
>> http://mailman.uib.no/listinfo/corpora
> 
> _______________________________________________
> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20130909/07304753/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora