I have checked some of the tools which are provided by some Indian universities but found some short stoppers
http://nlp.amrita.edu:8080/NLP4U/download.html , the links are broken and cannot access any of their NLP tools, i did try to contact the university.
http://au-kbc.org/nlp/corpusrelease.html , needs to approve your request to download it, unfortunately I wasnt still able to get in.
https://tdil-dc.in/index.php?option=com_download&task=showresourceDetails&toolid=356&lang=en , all their tools require a linux-based system and their downloads not working.
So far i've narrowed it down to two tools that i think, it may work
http://ltrc.iiit.ac.in/showfile.php?filename=downloads/shallow_parser.php, this seems like a good option, but there are no instructions on how to work with it. Has anyone tried this before?
https://github.com/datquocnguyen/RDRPOSTagger, provides only Universal POS annotation for Tamil
Are there any other resources that anyone would recommend me to look into? Or any resources to learn how to train a tagger with a POS tagged corpus would also be highly appreciated.
Thank you for all your help!
Hi there, I am using NLTK library. You can download it online from Standard University website, otherwise, please go through this article. I hope it can help
http://shodhganga.inflibnet.ac.in/bitstream/10603/34235/12/12_chapter%207.pdf
Chakir Mahjoubi Hi, thanks for that article! But NLTK does not provide support for Tamil language yet right?
Answering my own question, I was able to find a python library which does POS tagging for Tamil. But the limitation is that, it can tag only Universal POS annotation (https://universaldependencies.org/u/pos/all.html) . This is the only library available openly for Tamil POS tagging so far.
RDRPOSTagger: https://github.com/datquocnguyen/RDRPOSTagger, this is an R open source package.
The same package has been forked and modified to be used in python as RippleTagger: https://github.com/EmilStenstrom/rippletagger with an accuracy of 82% for Tamil.
Could you please check those links:
- https://mailman.uib.no/public/corpora/2016-September/025219.html
- https://www.amrita.edu/publication/hierarchal-pos-tagging-tamil-language-using-machine-learning-approach
I hope this could help. For Python, I suggest that you use version 4.0. It is I think the one that support NLTK.
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning