0

What are some of the stopwords in Tamil. The ones I could come up with were 'வது', 'ஆம்'.

  • tdil-dc.in/… has a downloadable list of what is alleged to be 185 stop words for Tamil. I cannot attest to its quality, but the organization seems vaguely credible. The download link requires you to register; I have not attempted to proceed past that point. – tripleee Jun 12 '18 at 10:01
  • Ah, I have tried this before. This organization requires you to send a fax. – sudha Jun 13 '18 at 5:40
1

You can look for the most frequent words in Tamil and make, e.g., the 50 most frequent words your stopwords.

List of the 50 most frequent words for many languages are available from the Wortschatz portal at Leipzig University, here is a direct links to 2017's TOP 50 words for Tamil.

| improve this answer | |
  • This seems a bit crude. Doing some statistics on the frequency distribution and/or checking with an informant who can read the language seem like obvious and not particularly sophisticated improvement ideas, though it is of course possible that they won't change the end result. 50 seems pretty low for a completely arbitrary number to pick for this purpose. – tripleee Mar 15 '18 at 17:51
  • I agree with @tripleee the list contains words for government and the name of a city. – sudha Jun 11 '18 at 13:13
  • The Wortschatz collection happens to make the top-50 available as a separate download by default; but if you speak the language, you could probably download the full corpus and pick out good candidates from something similar (top-200, top-500? This also depends on orthographical conventions of the language etc). – tripleee Jun 11 '18 at 13:45
  • @tripleee thank you! I am trying to do something of that sort. Do you think using a tf-idf score would help? – sudha Jun 12 '18 at 12:33
  • Just the IDF is probably sufficient already actually. – tripleee Jun 12 '18 at 13:21

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.