2
$\begingroup$

I am trying to tag Hindi text using python(nltk library). I have been successful but I am unable to understand some of the tags.

I tried searching for tagset but the only information I could find is about some upenn_tagset. The tags that I get matches only some of the tags from upenn_tagset.

Does anyone know how to get the tagset for hindi?

$\endgroup$
1
  • $\begingroup$ Can you share the specific tags? $\endgroup$ Dec 18 '19 at 19:43
1
$\begingroup$

I'm not familiar with NLTK but the tagset must come from the annotated corpus which was used to train the tagger. According to https://www.nltk.org/book/ch05.html (section 2.2), if you can find the name of the Hindi corpus you should be able to access the words with their tags with:

nltk.corpus.<corpus>.tagged_words()

Apparently this is the list of corpora available in NLTK: http://www.nltk.org/nltk_data/, it might help finding the one used to train the Hindi POS tagger.

Once you identify the corpus, it's likely that you can find explanations about the tagset and annotation process by searching for papers about its creation by the original authors.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.