Friday, April 24, 2009

[research] concept based searching in Tamil

this is pretty young to tamil research.
you google for மும்பை தங்கும் வசதி and you get 780 useless documents.
google expects you to be more intelligent than just knowing how to type in tamil.
who tells a search engine that மும்பை, பம்பாய், பாம்பே, மும்பே all refers to the city of mumbai.
how can a search engine understand that தங்கும் வசதி refers to motel.
concept based search aims at 'understanding' the web.

i am working with a small team here.
we are trying our best to make a search engine 'understand' tourism documents in tamil.
hoping to extend it to other domains.

go to powerset.com
type 'who directed roja?'
explore to get a feel of concept based search.

a long way to go....

1 comment:

sandoz said...

http://www.dosai.com/

This works decent with single search word.. No clue how it works though..

Interestingly when I searched மும்பை தங்கும் வசதி your blog was the first result!!